cjw/amd-r9700-ai-toolboxes
1
0
Fork 0
Code Issues Pull Requests Actions 6 Packages Projects Releases Wiki Activity

added initial benchmark results

Browse Source
This commit is contained in:
Donato Capitella
2025-11-30 11:17:29 +00:00
parent 649cf0ab45
commit 1b8bfa7230
199 changed files with 11788 additions and 44691 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7-nightly-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1543.57 ± 178.00 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.52 ± 0.14 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7-nightly-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 655.20 ± 11.76 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 52.00 ± 0.03 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7-nightly__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1626.90 ± 24.29 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 56.95 ± 0.63 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7-nightly__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 645.15 ± 5.88 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 51.85 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7.9-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1573.70 ± 86.22 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.94 ± 0.15 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7.9-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 681.11 ± 9.98 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.99 ± 0.09 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7.9__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1617.14 ± 72.83 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.48 ± 0.08 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm-7.9__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 671.33 ± 34.85 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.44 ± 0.08 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm6_4_4-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1699.96 ± 17.30 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 58.28 ± 0.14 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm6_4_4-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 672.15 ± 7.22 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 52.41 ± 0.51 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm6_4_4__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1532.28 ± 244.19 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.91 ± 0.12 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm6_4_4__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 684.36 ± 9.67 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 52.62 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm7.1-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1590.83 ± 63.43 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.90 ± 0.10 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm7.1-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 683.23 ± 3.48 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.93 ± 0.07 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm7.1__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1577.37 ± 164.50 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 58.25 ± 0.10 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__rocm7.1__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 689.58 ± 10.51 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 57.45 ± 0.06 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__vulkan_amdvlk__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Ryzen 9 9900X3D 12-Core Processor (AMD open-source driver) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 32768 | int dot: 1 | matrix cores: none
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 336.52 ± 4.47 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 14.83 ± 0.41 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-BF16-00001-of-00002__vulkan_radv__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Ryzen 9 9900X3D 12-Core Processor (RADV RAPHAEL_MENDOCINO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 65536 | int dot: 1 | matrix cores: none
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 1685.39 ± 159.20 |
| qwen3moe 30B.A3B BF16 | 56.89 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 25.37 ± 0.82 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7-nightly-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3029.83 ± 29.89 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 107.69 ± 0.79 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7-nightly-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3014.05 ± 33.10 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 108.03 ± 0.41 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7-nightly__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2974.25 ± 26.95 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 107.14 ± 0.45 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7-nightly__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2905.24 ± 49.34 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 107.17 ± 0.45 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7.9-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2819.42 ± 24.05 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 101.46 ± 0.29 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7.9-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2831.10 ± 38.83 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 101.12 ± 0.38 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7.9__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2900.63 ± 29.61 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 102.39 ± 0.13 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm-7.9__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2864.22 ± 14.47 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 102.81 ± 0.37 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm6_4_4-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2819.84 ± 23.93 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 106.53 ± 0.43 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm6_4_4-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2841.40 ± 38.04 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 106.60 ± 0.31 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm6_4_4__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2945.06 ± 35.71 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 105.64 ± 0.54 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm6_4_4__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2958.74 ± 12.12 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 105.96 ± 0.24 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm7.1-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2799.37 ± 7.73 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 101.37 ± 0.25 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm7.1-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2816.37 ± 26.52 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 101.43 ± 0.23 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm7.1__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2873.53 ± 17.86 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 103.10 ± 0.46 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__rocm7.1__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2876.61 ± 20.40 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 102.77 ± 0.31 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__vulkan_amdvlk__fa1__single.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Ryzen 9 9900X3D 12-Core Processor (AMD open-source driver) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 32768 | int dot: 1 | matrix cores: none
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 1532.93 ± 158.64 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 112.70 ± 0.06 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-Q4_K_M__vulkan_radv__fa1__single.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Ryzen 9 9900X3D 12-Core Processor (RADV RAPHAEL_MENDOCINO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 65536 | int dot: 1 | matrix cores: none
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 1353.29 ± 143.93 |
| qwen3moe 30B.A3B Q4_K - Medium | 17.35 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 102.36 ± 0.73 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7-nightly-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 746.19 ± 3.50 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 31.19 ± 0.03 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7-nightly-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 691.91 ± 4.04 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 28.64 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7-nightly__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 744.87 ± 4.36 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.61 ± 0.95 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7-nightly__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 689.70 ± 3.91 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 28.70 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7.9-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 734.30 ± 4.45 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.83 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7.9-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 755.65 ± 2.29 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 31.06 ± 0.08 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7.9__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 738.77 ± 1.95 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.85 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm-7.9__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 761.15 ± 0.65 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.97 ± 0.07 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm6_4_4-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 678.02 ± 26.20 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.84 ± 0.12 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm6_4_4-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 685.55 ± 6.26 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 27.85 ± 0.17 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm6_4_4__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 664.88 ± 136.58 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 31.27 ± 0.03 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm6_4_4__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 692.42 ± 0.22 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 27.42 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm7.1-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 723.69 ± 8.04 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.85 ± 0.03 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm7.1-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 751.01 ± 2.42 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 31.17 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm7.1__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 731.65 ± 3.93 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 30.99 ± 0.04 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__rocm7.1__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | pp512 | 755.98 ± 3.76 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | ROCm | 99 | 1 | tg128 | 31.20 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__vulkan_amdvlk__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Ryzen 9 9900X3D 12-Core Processor (AMD open-source driver) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 32768 | int dot: 1 | matrix cores: none
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | Vulkan | 99 | 1 | pp512 | 358.44 ± 0.84 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | Vulkan | 99 | 1 | tg128 | 30.61 ± 0.17 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Next-80B-A3B-Instruct-UD-Q4_K_XL__vulkan_radv__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Ryzen 9 9900X3D 12-Core Processor (RADV RAPHAEL_MENDOCINO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 65536 | int dot: 1 | matrix cores: none
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | Vulkan | 99 | 1 | pp512 | 295.49 ± 0.36 |
| qwen3next ?B Q4_K - Medium | 42.01 GiB | 79.67 B | Vulkan | 99 | 1 | tg128 | 21.56 ± 0.16 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7-nightly-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1475.87 ± 6.84 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.94 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7-nightly-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1209.56 ± 3.22 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.93 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7-nightly__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1470.53 ± 4.40 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.24 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7-nightly__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1194.33 ± 0.71 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.25 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7.9-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2639.48 ± 23.85 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.98 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7.9-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1841.65 ± 6.92 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.96 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7.9__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2781.05 ± 5.48 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.25 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm-7.9__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1905.36 ± 8.03 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.25 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm6_4_4-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2757.27 ± 20.99 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.01 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm6_4_4-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1924.47 ± 6.43 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.97 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm6_4_4__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2875.43 ± 26.48 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.26 ± 0.03 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm6_4_4__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1976.49 ± 8.76 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.27 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm7.1-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2620.76 ± 8.66 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.95 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm7.1-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1828.43 ± 7.53 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 30.97 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm7.1__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 2788.00 ± 4.06 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.29 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__rocm7.1__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | pp512 | 1912.89 ± 8.88 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | ROCm | 99 | 1 | tg128 | 31.30 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__vulkan_amdvlk__fa1__single.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Ryzen 9 9900X3D 12-Core Processor (AMD open-source driver) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 32768 | int dot: 1 | matrix cores: none
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | Vulkan | 99 | 1 | pp512 | 1502.24 ± 99.88 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | Vulkan | 99 | 1 | tg128 | 30.58 ± 0.21 |
build: fa0465954 (7205)
benchmark/results/gemma-3-12b-it-UD-Q8_K_XL__vulkan_radv__fa1__single.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Ryzen 9 9900X3D 12-Core Processor (RADV RAPHAEL_MENDOCINO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 65536 | int dot: 1 | matrix cores: none
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | Vulkan | 99 | 1 | pp512 | 1328.74 ± 100.82 |
| gemma3 12B Q8_0 | 13.40 GiB | 11.77 B | Vulkan | 99 | 1 | tg128 | 27.25 ± 0.14 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7-nightly-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1590.78 ± 0.99 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.45 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7-nightly-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 183.73 ± 0.09 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.30 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7-nightly__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1522.51 ± 5.17 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.46 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7-nightly__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 182.11 ± 0.08 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.32 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7.9-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1589.12 ± 0.47 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.48 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7.9-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 176.22 ± 0.11 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.49 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7.9__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1578.72 ± 2.69 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.48 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm-7.9__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 175.03 ± 0.20 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.48 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm6_4_4-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1270.25 ± 1.64 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.49 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm6_4_4-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 178.54 ± 0.09 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.36 ± 0.01 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm6_4_4__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1313.33 ± 2.59 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.50 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm6_4_4__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 180.29 ± 0.06 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.38 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm7.1-rocwmma__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1559.16 ± 2.09 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.48 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm7.1-rocwmma__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 174.47 ± 0.15 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.48 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm7.1__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 1586.67 ± 5.72 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.49 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__rocm7.1__hblt0__fa1__dual.log
+11
View File
@@ -0,0 +1,11 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 2 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
Device 1: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | pp512 | 175.85 ± 0.19 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | ROCm | 99 | 1 | tg128 | 9.49 ± 0.00 |
build: fa0465954 (7205)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__vulkan_amdvlk__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (AMD open-source driver) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 32768 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Ryzen 9 9900X3D 12-Core Processor (AMD open-source driver) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 32768 | int dot: 1 | matrix cores: none
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
ggml_vulkan: Device memory allocation of size 2819260416 failed.
ggml_vulkan: Requested buffer size exceeds device buffer size limit: ErrorOutOfDeviceMemory
main: error: failed to load model '/home/kyuz0/models/gemma-3/BF16/gemma-3-27b-it-BF16-00001-of-00002.gguf'
✖ ! [vulkan_amdvlk] gemma-3-27b-it-BF16-00001-of-00002__fa1 failed (exit 0)
benchmark/results/gemma-3-27b-it-BF16-00001-of-00002__vulkan_radv__fa1__dual.log
+12
View File
@@ -0,0 +1,12 @@
WARNING: radv is not a conformant Vulkan implementation, testing use only.
WARNING: radv is not a conformant Vulkan implementation, testing use only.
ggml_vulkan: Found 3 Vulkan devices:
ggml_vulkan: 0 = AMD Ryzen 9 9900X3D 12-Core Processor (RADV RAPHAEL_MENDOCINO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 32 | shared memory: 65536 | int dot: 1 | matrix cores: none
ggml_vulkan: 1 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
ggml_vulkan: 2 = AMD Radeon AI PRO R9700 (RADV GFX1201) (radv) | uma: 0 | fp16: 1 | bf16: 1 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | Vulkan | 99 | 1 | pp512 | 557.39 ± 0.99 |
| gemma3 27B BF16 | 50.31 GiB | 27.01 B | Vulkan | 99 | 1 | tg128 | 9.08 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7-nightly-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6665.84 ± 134.80 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 123.76 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7-nightly-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6646.29 ± 124.81 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 123.48 ± 0.07 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7-nightly__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6415.41 ± 135.85 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 127.03 ± 0.04 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7-nightly__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6220.27 ± 379.38 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 126.86 ± 0.05 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7.9-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6010.31 ± 117.03 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 121.51 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7.9-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6011.29 ± 107.57 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 121.44 ± 0.04 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7.9__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6267.43 ± 120.91 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 124.50 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm-7.9__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon Graphics, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6267.38 ± 130.90 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 124.27 ± 0.02 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm6_4_4-rocwmma__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6289.53 ± 138.61 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 122.62 ± 0.08 |
build: fa0465954 (7205)
benchmark/results/gemma-3-4b-it-Q3_K_S__rocm6_4_4-rocwmma__hblt0__fa1__single.log
+10
View File
@@ -0,0 +1,10 @@
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 ROCm devices:
Device 0: AMD Radeon AI PRO R9700, gfx1201 (0x1201), VMM: no, Wave Size: 32
| model | size | params | backend | ngl | fa | test | t/s |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | pp512 | 6305.84 ± 144.66 |
| gemma3 4B Q3_K - Small | 1.80 GiB | 3.88 B | ROCm | 99 | 1 | tg128 | 122.80 ± 0.04 |
build: fa0465954 (7205)

Some files were not shown because too many files have changed in this diff Show More