diff options
| author | Andrew M. B. Boktor <[email protected]> | 2012-04-18 16:33:33 -0800 |
|---|---|---|
| committer | Andrew Boktor <[email protected]> | 2014-08-14 13:19:05 -0700 |
| commit | 0cb76063ce079a0606598e145da351f134235ab5 (patch) | |
| tree | c1d2de65b2d03d4ef4257ec15037b656a03657c5 /configs | |
| parent | 5a65fb7a5f0cb1a8d7831bae33749e8bd15a947d (diff) | |
This changelist implements the following:
1. Adds support for using cuobjdump for both ptx and ptxplus execution. This has been tested with CUDA 4.0
. Ptxplus is no longer supported through decuda/decuda_to_ptxplus
2. Adds support for converting the SASS output by cuobjdump to ptxplus. This has been tested with CUDA 4.0
. The old path that extracts ptx from cubin files is still preserved
3. Adds a bank group model. (WARNING: memory config has changed, please adapt yours). To disable the bank groups model, set nbkgrp to 1 and tCCDL and tRTPL to 0
Diff the configuration files to learn about how to use those new options.
Merging
//depot/gpgpu_sim_research/fermi-test/distribution/...
to //depot/gpgpu_sim_research/fermi/distribution/...
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 12023]
Diffstat (limited to 'configs')
| -rw-r--r-- | configs/Fermi/gpgpusim.config | 14 | ||||
| -rw-r--r-- | configs/QuadroFX5800/gpgpusim.config | 15 |
2 files changed, 17 insertions, 12 deletions
diff --git a/configs/Fermi/gpgpusim.config b/configs/Fermi/gpgpusim.config index b20fe86..e1bcf1c 100644 --- a/configs/Fermi/gpgpusim.config +++ b/configs/Fermi/gpgpusim.config @@ -66,8 +66,9 @@ -gpgpu_mem_addr_mapping dramid@8;00000000.00000000.00000000.00000000.0000RRRR.RRRRRRRR.RRBBBCCC.CCCSSSSS # GDDR5 timing from hynix H5GQ1H24AFR -# {nbk:tCCD:tRRD:tRCD:tRAS:tRP:tRC:CL:WL:tCDLR:tWR} --gpgpu_dram_timing_opt 16:2:5:12:28:12:35:10:7:6:12 +# to disable bank groups, set nbkgrp to 1 and tCCDL and tRTPL to 0 +# {nbk:nbkgrp:tCCD:tCCDL:tRRD:tRCD:tRAS:tRP:tRTPL:tRC:CL:WL:tCDLR:tWR} +-gpgpu_dram_timing_opt 16:4:2:3:5:12:28:12:2:35:10:7:6:12 # GDDR3 #-gpgpu_dram_timing_opt 8:2:8:12:25:10:35:10:7:6:11 @@ -80,9 +81,12 @@ -gpgpu_runtime_stat 500 -enable_ptx_file_line_stats 1 -# enable SASS execution -#-gpgpu_ptx_convert_to_ptxplus 1 -#-gpgpu_ptx_save_converted_ptxplus 1 +# Using cuobjdump to extract ptx/SASS +-gpgpu_ptx_use_cuobjdump 0 + +# SASS execution (only supported with CUDA >= 4.0) +-gpgpu_ptx_convert_to_ptxplus 0 +-gpgpu_ptx_save_converted_ptxplus 0 # enable operand collector -gpgpu_operand_collector_num_units_sp 6 diff --git a/configs/QuadroFX5800/gpgpusim.config b/configs/QuadroFX5800/gpgpusim.config index b5e1f01..6c33a85 100644 --- a/configs/QuadroFX5800/gpgpusim.config +++ b/configs/QuadroFX5800/gpgpusim.config @@ -37,19 +37,20 @@ -gpgpu_mem_address_mask 1 -gpgpu_mem_addr_mapping dramid@8;00000000.00000000.00000000.00000000.0000RRRR.RRRRRRRR.RRBBBCCC.CCCSSSSS # GDDR3 timing from Samsung K4J52324QH-HC12 @ 800MHz -# {nbk:tCCD:tRRD:tRCD:tRAS:tRP:tRC:CL:WL:tCDLR:tWR} --gpgpu_dram_timing_opt 8:2:8:12:25:10:35:10:7:6:11 +# {nbk:nbkgrp:tCCD:tCCDL:tRRD:tRCD:tRAS:tRP:tRTPL:tRC:CL:WL:tCDLR:tWR} +-gpgpu_dram_timing_opt 8:1:2:0:8:12:25:10:0:35:10:7:6:11 # stat collection -gpgpu_memlatency_stat 14 -gpgpu_runtime_stat 500 -enable_ptx_file_line_stats 1 -# enable SASS execution -# For this to work correctly, you need to build the binary using older cuda (<3.0) -# Make sure this condition is met before enabling ptxplus -#-gpgpu_ptx_convert_to_ptxplus 1 -#-gpgpu_ptx_save_converted_ptxplus 1 +# Using cuobjdump to extract ptx/SASS +-gpgpu_ptx_use_cuobjdump 0 + +# SASS execution (only supported with CUDA >= 4.0) +-gpgpu_ptx_convert_to_ptxplus 0 +-gpgpu_ptx_save_converted_ptxplus 0 # enable operand collector -gpgpu_operand_collector_num_units_sp 6 |