diff --git a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log index 1ef8133..a0cb708 100644 --- a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log +++ b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | ---: | --------------: | -------------------: | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | 0 | pp512 | 283.71 ± 2.88 | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | 0 | tg128 | 52.21 ± 0.00 | +| model | size | params | backend | ngl | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | pp512 | 674.34 ± 3.87 | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | tg128 | 52.10 ± 0.00 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log index 2e840f6..5cd16ba 100644 --- a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log +++ b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | n_ubatch | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | ---: | --------------: | -------------------: | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | 0 | pp2048 @ d32768 | 207.09 ± 0.00 | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | 0 | tg32 @ d32768 | 36.30 ± 0.00 | +| model | size | params | backend | ngl | n_ubatch | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | --------------: | -------------------: | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | pp2048 @ d32768 | 352.35 ± 0.56 | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | tg32 @ d32768 | 40.44 ± 0.01 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log index 679ec4a..8898523 100644 --- a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log +++ b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | ---: | --------------: | -------------------: | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | 0 | pp512 | 284.30 ± 2.38 | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | 0 | tg128 | 52.16 ± 0.00 | +| model | size | params | backend | ngl | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | pp512 | 679.62 ± 5.41 | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 1 | tg128 | 52.11 ± 0.00 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log index 00bc827..75ba15a 100644 --- a/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log +++ b/benchmark/results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | n_ubatch | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | ---: | --------------: | -------------------: | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | 0 | pp2048 @ d32768 | 198.59 ± 0.00 | -| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | 0 | tg32 @ d32768 | 36.34 ± 0.00 | +| model | size | params | backend | ngl | n_ubatch | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | --------------: | -------------------: | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | pp2048 @ d32768 | 352.62 ± 1.18 | +| gpt-oss 120B MXFP4 MoE | 59.02 GiB | 116.83 B | ROCm | 99 | 2048 | 1 | tg32 @ d32768 | 40.46 ± 0.03 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log index 16a54d1..e0b5cee 100644 --- a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log +++ b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | ---: | --------------: | -------------------: | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | 0 | pp512 | 816.54 ± 1.21 | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | 0 | tg128 | 73.67 ± 0.01 | +| model | size | params | backend | ngl | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | pp512 | 1734.91 ± 19.75 | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | tg128 | 73.48 ± 0.01 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log index 8667197..e5b1d94 100644 --- a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log +++ b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | n_ubatch | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | ---: | --------------: | -------------------: | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | 0 | pp2048 @ d32768 | 374.10 ± 0.00 | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | 0 | tg32 @ d32768 | 51.98 ± 0.00 | +| model | size | params | backend | ngl | n_ubatch | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | --------------: | -------------------: | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | pp2048 @ d32768 | 557.68 ± 0.88 | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | tg32 @ d32768 | 57.80 ± 0.01 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log index 12ce75a..5330505 100644 --- a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log +++ b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | ---: | --------------: | -------------------: | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | 0 | pp512 | 821.55 ± 13.20 | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | 0 | tg128 | 73.73 ± 0.01 | +| model | size | params | backend | ngl | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | pp512 | 1709.32 ± 28.51 | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 1 | tg128 | 73.53 ± 0.00 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log index 4c7fde9..52dd5fe 100644 --- a/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log +++ b/benchmark/results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log @@ -1,8 +1,8 @@ ggml_cuda_init: found 1 ROCm devices: Device 0: Radeon 8060S Graphics, gfx1151 (0x1151), VMM: no, Wave Size: 32 -| model | size | params | backend | ngl | n_ubatch | fa | mmap | test | t/s | -| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | ---: | --------------: | -------------------: | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | 0 | pp2048 @ d32768 | 355.59 ± 0.00 | -| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | 0 | tg32 @ d32768 | 52.02 ± 0.00 | +| model | size | params | backend | ngl | n_ubatch | fa | test | t/s | +| ------------------------------ | ---------: | ---------: | ---------- | --: | -------: | -: | --------------: | -------------------: | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | pp2048 @ d32768 | 541.93 ± 1.58 | +| gpt-oss 20B MXFP4 MoE | 11.27 GiB | 20.91 B | ROCm | 99 | 2048 | 1 | tg32 @ d32768 | 57.84 ± 0.02 | -build: 9c142e3a2 (7670) +build: e0c93af2a (7938) diff --git a/docs/results.json b/docs/results.json index 824d7dd..e087ac3 100644 --- a/docs/results.json +++ b/docs/results.json @@ -1,6 +1,6 @@ { "meta": { - "generated_at": "2026-01-23T15:10:46Z", + "generated_at": "2026-02-04T17:41:39Z", "system_info": { "distro": "Fedora Linux 42 (Workstation Edition)", "kernel": "6.18.3-100.fc42.x86_64", @@ -19,6 +19,10 @@ { "hash": "a14b960bc", "number": "7816" + }, + { + "hash": "e0c93af2a", + "number": "7938" } ], "environments": [ @@ -16138,13 +16142,13 @@ "context": "default", "context_tokens": null, "test": "pp512", - "tps_mean": 283.71, - "tps_std": 2.88, + "tps_mean": 674.34, + "tps_std": 3.87, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16152,8 +16156,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16166,13 +16170,13 @@ "context": "default", "context_tokens": null, "test": "tg128", - "tps_mean": 52.21, + "tps_mean": 52.1, "tps_std": 0.0, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16180,8 +16184,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16194,13 +16198,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "pp2048 @ d32768", - "tps_mean": 207.09, - "tps_std": 0.0, + "tps_mean": 352.35, + "tps_std": 0.56, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16208,8 +16212,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16222,13 +16226,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "tg32 @ d32768", - "tps_mean": 36.3, - "tps_std": 0.0, + "tps_mean": 40.44, + "tps_std": 0.01, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16236,8 +16240,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16250,13 +16254,13 @@ "context": "default", "context_tokens": null, "test": "pp512", - "tps_mean": 284.3, - "tps_std": 2.38, + "tps_mean": 679.62, + "tps_std": 5.41, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16264,8 +16268,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16278,13 +16282,13 @@ "context": "default", "context_tokens": null, "test": "tg128", - "tps_mean": 52.16, + "tps_mean": 52.11, "tps_std": 0.0, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16292,8 +16296,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16306,13 +16310,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "pp2048 @ d32768", - "tps_mean": 198.59, - "tps_std": 0.0, + "tps_mean": 352.62, + "tps_std": 1.18, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16320,8 +16324,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -16334,13 +16338,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "tg32 @ d32768", - "tps_mean": 36.34, - "tps_std": 0.0, + "tps_mean": 40.46, + "tps_std": 0.03, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 116.83, "file_size_gib": 59.02, "name_params_b": 116.83, @@ -16348,8 +16352,8 @@ "log": "results/gpt-oss-120b-mxfp4-00001-of-00003__rocm7-nightlies__hblt0__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17258,13 +17262,13 @@ "context": "default", "context_tokens": null, "test": "pp512", - "tps_mean": 816.54, - "tps_std": 1.21, + "tps_mean": 1734.91, + "tps_std": 19.75, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17272,8 +17276,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17286,13 +17290,13 @@ "context": "default", "context_tokens": null, "test": "tg128", - "tps_mean": 73.67, + "tps_mean": 73.48, "tps_std": 0.01, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17300,8 +17304,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17314,13 +17318,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "pp2048 @ d32768", - "tps_mean": 374.1, - "tps_std": 0.0, + "tps_mean": 557.68, + "tps_std": 0.88, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17328,8 +17332,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17342,13 +17346,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "tg32 @ d32768", - "tps_mean": 51.98, - "tps_std": 0.0, + "tps_mean": 57.8, + "tps_std": 0.01, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17356,8 +17360,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17370,13 +17374,13 @@ "context": "default", "context_tokens": null, "test": "pp512", - "tps_mean": 821.55, - "tps_std": 13.2, + "tps_mean": 1709.32, + "tps_std": 28.51, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17384,8 +17388,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17398,13 +17402,13 @@ "context": "default", "context_tokens": null, "test": "tg128", - "tps_mean": 73.73, - "tps_std": 0.01, + "tps_mean": 73.53, + "tps_std": 0.0, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17412,8 +17416,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17426,13 +17430,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "pp2048 @ d32768", - "tps_mean": 355.59, - "tps_std": 0.0, + "tps_mean": 541.93, + "tps_std": 1.58, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17440,8 +17444,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, { @@ -17454,13 +17458,13 @@ "context": "longctx32768", "context_tokens": 32768, "test": "tg32 @ d32768", - "tps_mean": 52.02, - "tps_std": 0.0, + "tps_mean": 57.84, + "tps_std": 0.02, "error": false, "error_type": null, "backend": "ROCm", "ngl": 99, - "mmap": 0, + "mmap": null, "params_b": 20.91, "file_size_gib": 11.27, "name_params_b": 20.91, @@ -17468,8 +17472,8 @@ "log": "results/gpt-oss-20b-mxfp4__rocm7-nightlies__hblt0__fa1__longctx32768.log", "rpc": false, "build": { - "hash": "9c142e3a2", - "number": "7670" + "hash": "e0c93af2a", + "number": "7938" } }, {