[hipblaslt] added problem sizes for PTS #2560
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,17 @@ | ||
| --- | ||
| include: ../../../tests/data/hipblaslt_common.yaml | ||
| include: ../../../tests/data/matmul_common.yaml | ||
|
|
||
| Tests: | ||
| - name: matmul_example | ||
| function: | ||
| matmul: *real_precisions | ||
| M: 128 | ||
| N: 128 | ||
| K: 128 | ||
| alpha_beta: *alpha_beta_range | ||
| transA_transB: *transA_transB_range | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
| ... |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,74 @@ | ||
| --- | ||
| include: ../../../tests/data/hipblaslt_common.yaml | ||
| include: ../../../tests/data/matmul_common.yaml | ||
|
|
||
| Definitions: | ||
| Tests: | ||
| - name: hpa_half_precision_rcmt_mphb_nn | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 8192, N: 320, K: 320 } | ||
| - { M: 2048, N: 640, K: 640 } | ||
| - { M: 512, N: 1280, K: 1280 } | ||
| - { M: 8192, N: 320 , K: 320 } | ||
| - { M: 512, N: 10240, K: 1280 } | ||
| - { M: 2048, N: 5120, K: 640 } | ||
| - { M: 8192, N: 2560, K: 320 } | ||
| - { M: 512, N: 1280, K: 5120 } | ||
| - { M: 2048, N: 640, K: 2560 } | ||
| - { M: 154, N: 320, K: 768 } | ||
| - { M: 154, N: 1280, K: 768 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: hpa_half_precision_rcmt_mphb_b_nn | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 4096, N: 40, K: 4096 } | ||
| - { M: 1024, N: 80, K: 1024 } | ||
| - { M: 1024, N: 88, K: 77 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| batch_count: 16 | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: hpa_half_precision_rcmt_mphb_b_nt | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 4096, N: 4096, K: 40 } | ||
| - { M: 1024, N: 1024, K: 80 } | ||
| - { M: 4096, N: 77, K: 40 } | ||
| - { M: 256, N: 77, K: 160 } | ||
| - { M: 1024, N: 77, K: 160 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: T | ||
| batch_count: 16 | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: hpa_half_precision_rcmt_mhb_tn | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 10240, N: 16 , K: 64640 } | ||
| initialization: trig_float | ||
| transA: T | ||
| transB: N | ||
| splitk: 1 | ||
| iters: 1000 | ||
| cold_iters: 4000 | ||
| rotating: 512 | ||
| api_method: 2 | ||
solaslin marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| ... | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,165 @@ | ||
| --- | ||
| include: ../../../tests/data/hipblaslt_common.yaml | ||
| include: ../../../tests/data/matmul_common.yaml | ||
|
|
||
| Definitions: | ||
|
|
||
| test precision: &hpc_gemm_shapes | ||
| - *single_precision | ||
| - *hpa_bf16_precision | ||
| - *hpa_half_precision | ||
| - *f8_fnuz_precision_dst_f8_fnuz | ||
| - *f8_fnuz_precision_dst_fp16 | ||
|
|
||
|
Comment on lines
+1
to
+13
Contributor
Author
There was a problem hiding this comment. Message for Mahmood, Wasiq:
The two includes already define almost all commonly used datatypes, transA,B...etc. definitions which are also used for gtest. So adding the two line can avoid copying-pasting. But you still can add your own definition (usually is a customized lists) such as:
the definition from line#8 to line#12 can be found in the two included yamls, however, you can still make a new group for your own test cases. The rest of the content simply look like any one in matmul_gtest.yaml |
||
| Tests: | ||
| - name: hpa_half_precision_mad_summary | ||
| function: | ||
| matmul: *hpc_gemm_shapes | ||
| matrix_size: | ||
| - { M: 1, N: 10240, K: 8192 } | ||
| - { M: 1, N: 12288, K: 4096 } | ||
| - { M: 1, N: 12288, K: 12288 } | ||
| - { M: 1, N: 15360, K: 5120 } | ||
| - { M: 1, N: 16384, K: 2048 } | ||
| - { M: 1, N: 18432, K: 6144 } | ||
| - { M: 1, N: 20480, K: 5120 } | ||
| - { M: 1, N: 24576, K: 5120 } | ||
| - { M: 1, N: 25600, K: 6400 } | ||
| - { M: 1, N: 3072, K: 2048 } | ||
| - { M: 1, N: 3584, K: 3584 } | ||
| - { M: 1, N: 4096, K: 4096 } | ||
| - { M: 1, N: 4096, K: 11008 } | ||
| - { M: 1, N: 4096, K: 22016 } | ||
| - { M: 1, N: 4608, K: 3584 } | ||
| - { M: 1, N: 5120, K: 1536 } | ||
| - { M: 1, N: 5120, K: 12288 } | ||
| - { M: 1, N: 57344, K: 8192 } | ||
| - { M: 1, N: 6144, K: 4096 } | ||
| - { M: 1, N: 8192, K: 8192 } | ||
| - { M: 2, N: 3072, K: 5120 } | ||
| - { M: 2, N: 5120, K: 12288 } | ||
| - { M: 4, N: 12288, K: 4096 } | ||
| - { M: 4, N: 4096, K: 4096 } | ||
| - { M: 4, N: 5120, K: 12288 } | ||
| - { M: 4, N: 6144, K: 4096 } | ||
| - { M: 4, N: 2560, K: 10240 } | ||
| - { M: 4, N: 44032, K: 4096 } | ||
| - { M: 4, N: 5120, K: 1536 } | ||
| - { M: 8, N: 10240, K: 8192 } | ||
| - { M: 8, N: 11008, K: 4096 } | ||
| - { M: 8, N: 12288, K: 4096 } | ||
| - { M: 8, N: 15360, K: 5120 } | ||
| - { M: 8, N: 16384, K: 2048 } | ||
| - { M: 8, N: 18432, K: 6144 } | ||
| - { M: 8, N: 19200, K: 6400 } | ||
| - { M: 8, N: 20480, K: 2560 } | ||
| - { M: 8, N: 24576, K: 5120 } | ||
| - { M: 8, N: 2560, K: 2560 } | ||
| - { M: 8, N: 28672, K: 4096 } | ||
| - { M: 8, N: 3072, K: 3072 } | ||
| - { M: 8, N: 3584, K: 3584 } | ||
| - { M: 8, N: 3840, K: 2560 } | ||
| - { M: 8, N: 4096, K: 13696 } | ||
| - { M: 8, N: 4608, K: 4608 } | ||
| - { M: 8, N: 5120, K: 5120 } | ||
| - { M: 8, N: 5760, K: 5120 } | ||
| - { M: 8, N: 6144, K: 4096 } | ||
| - { M: 8, N: 7680, K: 5120 } | ||
| - { M: 8, N: 8192, K: 24576 } | ||
| - { M: 16, N: 5120, K: 1536 } | ||
| - { M: 32, N: 10240, K: 8192 } | ||
| - { M: 32, N: 11008, K: 4096 } | ||
| - { M: 32, N: 12288, K: 4096 } | ||
| - { M: 32, N: 128256,K: 4096 } | ||
| - { M: 32, N: 13696, K: 4096 } | ||
| - { M: 32, N: 15360, K: 5120 } | ||
| - { M: 32, N: 16384, K: 4096 } | ||
| - { M: 32, N: 18432, K: 6144 } | ||
| - { M: 32, N: 20480, K: 5120 } | ||
| - { M: 32, N: 24576, K: 6144 } | ||
| - { M: 32, N: 2560, K: 10240 } | ||
| - { M: 32, N: 28672, K: 4096 } | ||
| - { M: 32, N: 3072, K: 3072 } | ||
| - { M: 32, N: 4096, K: 14336 } | ||
| - { M: 32, N: 44032, K: 4096 } | ||
| - { M: 32, N: 49152, K: 4096 } | ||
| - { M: 32, N: 5120, K: 1536 } | ||
| - { M: 32, N: 5760, K: 5120 } | ||
| - { M: 32, N: 6144, K: 4096 } | ||
| - { M: 32, N: 7168, K: 5120 } | ||
| - { M: 32, N: 7680, K: 5120 } | ||
| - { M: 32, N: 8192, K: 24576 } | ||
| - { M: 48, N: 10240, K: 10240 } | ||
| - { M: 64, N: 5120, K: 1536 } | ||
| - { M: 76, N: 32768, K: 4096 } | ||
| - { M: 77, N: 12288, K: 4096 } | ||
| - { M: 96, N: 35840, K: 6656 } | ||
| - { M: 96, N: 6656, K: 6656 } | ||
| - { M: 128, N: 5120, K: 12288 } | ||
| - { M: 128, N: 8192, K: 6144 } | ||
| - { M: 154, N: 2304, K: 768 } | ||
| - { M: 154, N: 3072, K: 1024 } | ||
| - { M: 197, N: 2304, K: 768 } | ||
| - { M: 256, N: 3072, K: 5120 } | ||
| - { M: 308, N: 2304, K: 768 } | ||
| - { M: 384, N: 2304, K: 768 } | ||
| - { M: 512, N: 3072, K: 5120 } | ||
| - { M: 1024, N: 1024, K: 4608 } | ||
| - { M: 1024, N: 16384, K: 4096 } | ||
| - { M: 1024, N: 20480, K: 5120 } | ||
| - { M: 1024, N: 2560, K: 10240 } | ||
| - { M: 1024, N: 3072, K: 2048 } | ||
| - { M: 1024, N: 4096, K: 1024 } | ||
| - { M: 1024, N: 5120, K: 13696 } | ||
| - { M: 1024, N: 768, K: 768 } | ||
| - { M: 1232, N: 2304, K: 768 } | ||
| - { M: 1536, N: 3072, K: 1024 } | ||
| - { M: 2048, N: 2048, K: 8192 } | ||
| - { M: 2048, N: 21504, K: 7168 } | ||
| - { M: 2048, N: 27648, K: 5120 } | ||
| - { M: 2048, N: 3072, K: 8192 } | ||
| - { M: 2048, N: 5120, K: 13824 } | ||
| - { M: 2048, N: 768, K: 768 } | ||
| - { M: 2464, N: 2304, K: 768 } | ||
| - { M: 3072, N: 2560, K: 2560 } | ||
| - { M: 3072, N: 49152, K: 12288 } | ||
| - { M: 4096, N: 1024, K: 4096 } | ||
| - { M: 4096, N: 16384, K: 2048 } | ||
| - { M: 4096, N: 2560, K: 10240 } | ||
| - { M: 4096, N: 3072, K: 1024 } | ||
| - { M: 4096, N: 5120, K: 5120 } | ||
| - { M: 4096, N: 768, K: 768 } | ||
| - { M: 8192, N: 20480, K: 2560 } | ||
| - { M: 8192, N: 3072, K: 1024 } | ||
| - { M: 8192, N: 4096, K: 1024 } | ||
| - { M: 8192, N: 768, K: 768 } | ||
| - { M: 10240, N: 5120, K: 5120 } | ||
| - { M: 12288, N: 1024, K: 4096 } | ||
| - { M: 14336, N: 16384, K: 6144 } | ||
| - { M: 16384, N: 1024, K: 4096 } | ||
| - { M: 16384, N: 16384, K: 4096 } | ||
| - { M: 16384, N: 2560, K: 2560 } | ||
| - { M: 16384, N: 3072, K: 1024 } | ||
| - { M: 16384, N: 4096, K: 13696 } | ||
| - { M: 16384, N: 768, K: 768 } | ||
| - { M: 20480, N: 128, K: 1 } | ||
| - { M: 25216, N: 2304, K: 768 } | ||
| - { M: 32768, N: 2048, K: 2048 } | ||
| - { M: 32768, N: 4544, K: 4544 } | ||
| - { M: 32768, N: 768, K: 768 } | ||
| - { M: 49152, N: 10240, K: 10240 } | ||
| - { M: 57344, N: 10240, K: 8192 } | ||
| - { M: 65536, N: 3840, K: 1280 } | ||
| - { M: 65536, N: 768, K: 768 } | ||
| - { M: 73728, N: 2304, K: 768 } | ||
| - { M: 77824, N: 6656, K: 6656 } | ||
| - { M: 802816, N: 16, K: 32 } | ||
| - { M: 8000000, N: 16, K: 32 } | ||
| - { M: 10000000, N: 1, K: 2 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
| ... | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,41 @@ | ||
| --- | ||
| include: ../../../tests/data/hipblaslt_common.yaml | ||
| include: ../../../tests/data/matmul_common.yaml | ||
|
|
||
| Definitions: | ||
| Tests: | ||
| - name: test_rvs | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 1024, N: 1024, K: 1024 } | ||
| - { M: 2048, N: 2048, K: 2048 } | ||
| - { M: 3072, N: 3072, K: 3072 } | ||
| - { M: 4096, N: 8192, K: 4096 } | ||
| - { M: 4864, N: 8192, K: 4096 } | ||
| - { M: 8096, N: 8096, K: 8096 } | ||
| - { M: 8192, N: 8192, K: 8192 } | ||
| - { M: 8192, N: 16384, K: 8192 } | ||
| - { M: 8640, N: 8640, K: 8640 } | ||
| - { M: 16182, N: 16182, K: 16182 } | ||
| - { M: 28000, N: 28000, K: 28000 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: test_rvs_batched | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 8640, N: 8640, K: 8640 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| batch_count: 96 | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
| ... |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,86 @@ | ||
| --- | ||
| include: ../../../tests/data/hipblaslt_common.yaml | ||
| include: ../../../tests/data/matmul_common.yaml | ||
|
|
||
| Definitions: | ||
| Tests: | ||
| - name: real_precisions_agfhc | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 4096, N: 8192, K: 4096 } | ||
| - { M: 8192, N: 8192, K: 8192 } | ||
| - { M: 8192, N: 8192, K: 8960 } | ||
| - { M: 8192, N: 4096, K: 12288 } | ||
| - { M: 8192, N: 16384, K: 8192 } | ||
| - { M: 8640, N: 8640, K: 8640 } | ||
| - { M: 28000, N: 28000, K: 28000 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: real_precisions_agfhc_batched | ||
| function: | ||
| matmul: *hpa_half_precision | ||
| matrix_size: | ||
| - { M: 8640, N: 8640, K: 8640 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| batch_count: 96 | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: f8_fnuz_precision_dst_f8_fnuz_agfhc | ||
| function: | ||
| matmul: *f8_fnuz_precision_dst_f8_fnuz | ||
| matrix_size: | ||
| - { M: 8192, N: 8192, K: 8192 } | ||
| - { M: 8192, N: 16384,K: 8192 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: i8_precision_dst_i8_agfhc | ||
| function: | ||
| matmul: *i8_precision_dst_i8 | ||
| matrix_size: | ||
| - { M: 8192, N: 17792,K: 13312 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: mxf6_mxf6_precision_dst_fp32_agfhc | ||
| function: | ||
| matmul: *mxf6_mxf6_precision_dst_fp32 | ||
| matrix_size: | ||
| - { M: 2048, N: 2048,K: 2048 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
|
|
||
| - name: mxf4_mxf4_precision_dst_fp32_agfhc | ||
| function: | ||
| matmul: *mxf4_mxf4_precision_dst_fp32 | ||
| matrix_size: | ||
| - { M: 2048, N: 2048,K: 2048 } | ||
| initialization: trig_float | ||
| transA: N | ||
| transB: N | ||
| iters: 1000 | ||
| cold_iters: 1000 | ||
| rotating: 512 | ||
| ... |
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Uh oh!
There was an error while loading. Please reload this page.