aboutsummaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2016-08-05Deleted useless commentssspenst
2016-08-05Added ptx_warp_info to know how many threads within a warp have executedsspenst
2016-08-05bsmad gives the correct output in the small cases I have tried, still need ↵sspenst
to complete the TODOs noted in bsmad_impl
2016-08-04A thread executing BSMAD is now able to access information from all threads ↵sspenst
in its warp
2016-07-11Changed sst return value to be the address instead of index offsetsspenst
2016-07-11Reverted the previous commit to add a cleaner way of getting NUM_THREADS. ↵sspenst
Now, sst_impl doesn't functionally execute on the last indexed element of an array, but instead on the actual last thread that executes
2016-07-08Made gridDim and blockDim global variables so that they can be accessed from ↵sspenst
sst_impl
2016-07-08SST should now properly simulate the barrier operationsspenst
2016-07-07sst_impl cleanupsspenst
2016-07-07Indices are now stored corresponding to values. SST now returns the number ↵sspenst
of elements instead of the device memory address
2016-07-07SST instruction now returns the end address of the new sparse arraysspenst
2016-07-07SST instruction now updates the original array instead of storing the result ↵sspenst
in sstarr memory
2016-07-07Rough implementation of the SST instruction. It squeezes out the zeros that ↵sspenst
are in the sstarr memory and writes the data back into sstarr memory.
2016-07-06Added the ability to load from sstarr memory after data has been stored in itsspenst
2016-07-06Added sstarr memory, which works the same as shared memorysspenst
2016-07-06ADD: add knob to enable CDP in gpgpusim configJin Wang
2016-07-06MOD: modify to new structure name gpgpu_ptx_sim_infoJin Wang
2016-07-06BUG: extra bracketJin Wang
2016-07-06BUG: wrong declaration for m_args_aligned_sizeJin Wang
2016-07-06ADD: add kernel launching latency from stream to distributorJin Wang
2016-07-06ADD: print kernel parameter size footprint. BUG: concurrent kernels on same ↵Jin Wang
shader, should use hw_cta_id to store shared mem info
2016-07-06ADD: add stats for kernel launching and complete cycleJin Wang
2016-07-06BUG: concurrent kernels on same SM may occupy warps from running CTAsJin Wang
2016-07-06ADD: add separate cdp latencyJin Wang
2016-07-06ADD: add cdp latencyJin Wang
2016-07-06BUG: for concurrent kernels on same shader, should select kernel from the ↵Jin Wang
distributor directly
2016-07-06ADD: support concurrent kernels on one shaderJin Wang
2016-07-05Slight addition to the parser for CPTX aesthetics. You now begin and end ↵sspenst
your custom inserted PTX with CPTX_BEGIN and CPTX_END, respectively.
2016-07-05ADD: launch all device kernels at once in functional simulatorJin Wang
2016-07-05MOD: compute child parameter sizeJin Wang
2016-07-05MOD: schedule one child kernel each cycleJin Wang
2016-07-05ADD: delete streams created by cta when deleting kernelJin Wang
2016-07-05BUG: PTX section id. ADD: cudaDeviceSetLimit. BUG: parameter addresses for ↵Jin Wang
child kernels in CDP. BUG: .weak .entry and .weak .global directives in ptx file. BUG: empty_protected() for stream manager causes deadlock, change to empty()
2016-07-05BUG: kernels should return to stream if not pushed to concurrent kernel poolJin Wang
2016-07-05ADD: add support for cudaStreamCreateWithFlagsJin Wang
2016-07-05BUG: multiple child kernels finishJin Wang
2016-07-05BUG: do not handle cudaGetParameterBufferV2 and cudaLaunchDeviceV2 as ↵Jin Wang
call.uni in reconvergence
2016-07-05BUG: parameter alignmentJin Wang
2016-07-05MOD: add child kernel stream and scheduling supportJin Wang
2016-07-05ADD: add cudaGetParameterBufferV2 and add cudaLaunchDeviceV2 implementation. ↵Jin Wang
Kernel launch to stream not yet implemented
2016-07-05ADD: initial support for instruction group used by CDPJin Wang
2016-07-05ADD: handle child kernel name in mov instruction. ADD: detect call ↵Jin Wang
cudaGetParameterBufferV2 and call cudaLaunchDeviceV2
2016-07-05ADD: support ptxinfo for sm_35 and cuda 6.5Jin Wang
2016-07-04Initial SST recognition from PTX parsersspenst
2016-07-04Restored madp instruction.speverel
2016-07-04Reverted part of the previous commit so that our new changes related to DNNs ↵sspenst
can be done in a different branch
2016-06-16Added the ability to inject arbitrary PTX instructions. ↵speverel
This will be used to add custom instructions in the future; the imaginary instructions 'spr' and 'ama' have been added as samples.
2016-06-13If ptxas notices any duplicate errors, they now automatically get resolved ↵sspenst
and the program continues with the duplicate function/variable declarations removed.
2016-06-07The ptx parser now recognizes the NC option for ld.global, however this ↵sspenst
option is not actually implemented
2016-06-07Added support for cudaMemcpyDefault flag in cudaMemcpy. Also increased the ↵speverel
maximum allowable memory to 2GB and the compute version to 5.2.
generated by cgit v1.3 (git 2.53.0) at 2026-07-13 15:21:42 +0000