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

updated benchmarks

Browse Source
This commit is contained in:
Donato Capitella
2025-11-30 11:44:58 +00:00
parent 1b8bfa7230
commit 84d8d7ce5a
19 changed files with 1245 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1907.67 ± 21.27 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.81 ± 0.20 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1773.31 ± 14.12 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.79 ± 0.17 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1862.04 ± 18.24 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.40 ± 0.13 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 1742.37 ± 17.13 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.32 ± 0.18 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3031.95 ± 63.94 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.34 ± 0.30 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2751.33 ± 25.32 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 76.19 ± 1.96 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3071.16 ± 32.87 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 76.06 ± 0.28 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2770.42 ± 4.16 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 76.12 ± 0.30 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3092.59 ± 48.47 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 78.53 ± 0.37 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2729.27 ± 20.16 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 78.54 ± 0.36 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3189.75 ± 37.79 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.60 ± 0.37 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2814.49 ± 29.02 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.13 ± 1.33 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3029.45 ± 26.26 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.19 ± 0.28 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2730.97 ± 3.89 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 77.29 ± 0.29 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 3127.94 ± 27.77 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 76.22 ± 0.29 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | pp512 | 2799.47 ± 4.43 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | ROCm | 99 | 1 | tg128 | 76.11 ± 0.29 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 1886.26 ± 127.63 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 82.73 ± 0.05 |
build: fa0465954 (7205)
benchmark/results/Qwen3-Coder-30B-A3B-Instruct-UD-Q8_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 |
| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | Vulkan | 99 | 1 | pp512 | 1616.13 ± 164.97 |
| qwen3moe 30B.A3B Q8_0 | 33.51 GiB | 30.53 B | Vulkan | 99 | 1 | tg128 | 77.99 ± 0.17 |
build: fa0465954 (7205)
docs/results.json
+1045 -1
View File
File diff suppressed because it is too large Load Diff