667 Commits

Author SHA1 Message Date
Richard Yan
2bbee0542b update radiance main with ae changes 2025-01-31 22:26:44 -08:00
Richard Yan
ae552222d6 Merge branch 'asplos-ae' 2025-01-31 19:17:14 -08:00
Richard Yan
68e4ebb471 remove tensor dpu sv 2025-01-30 22:34:38 -08:00
Richard Yan
a88da88a63 fix microcode of manual job invocation 2025-01-29 01:17:24 -08:00
Hansung Kim
1cd1f6fdcd Bump vortex 2025-01-28 22:37:46 -08:00
Richard Yan
52eeed277b correct loop count to start after receiving command 2025-01-28 17:41:00 -08:00
Richard Yan
d38f69fc5e Merge branch 'main' of https://github.com/ucb-bar/radiance 2025-01-28 17:34:01 -08:00
Richard Yan
18064f0c3c update final config and connect completion io 2025-01-28 14:24:20 -08:00
Hansung Kim
17001efbf3 Remove emulator_generate and merge into emulator_tick 2025-01-05 00:04:28 -08:00
Hansung Kim
a4fa1522ab Add D data to DPI interface 2025-01-04 23:03:25 -08:00
Hansung Kim
e3080bf3ee Update DPI for tick/generate split 2024-12-04 18:07:55 -08:00
Hansung Kim
7cc40eedde Add EmulatorTile
also split core-specific config keys from radiance.memory to radiance.core.
2024-11-26 15:23:24 -08:00
Hansung Kim
33ff495feb Fix doc errors and warnings for memfuzzer 2024-11-26 11:49:26 -08:00
Hansung Kim
9d70370801 Fix deprecation warnings in FuzzerTile 2024-11-25 22:33:29 -08:00
Hansung Kim
3b71276c4a tensor: Do dot-product in fp16, only do accum in fp32
This is to better match Gemmini PEs doing MACs in full fp16, and only
doing accumulation in fp32.
2024-11-15 15:36:32 -08:00
Richard Yan
8c473f52e3 no tc client if not decoupled 2024-11-11 18:16:26 -08:00
Richard Yan
6488cb5c78 support no gemminis, trim debug prints 2024-11-11 16:14:02 -08:00
Hansung Kim
8056fa4ada Bump vortex 2024-11-09 22:35:40 -08:00
Hansung Kim
31f0905567 Add new opcode for compute and spad mvout 2024-11-09 22:35:14 -08:00
Richard Yan
c87858b6f4 new cisc operations 2024-11-08 20:49:49 -08:00
Richard Yan
0c6618c65d back to non-blocking read 2024-11-07 18:20:37 -08:00
Richard Yan
d49abf97ff amend config based on gemmini update 2024-11-03 21:07:06 -08:00
Richard Yan
9a5af03672 blocking gemmini fence and bump vortex 2024-11-01 02:45:08 -07:00
Hansung Kim
99da429cb1 tensor: Move C reg access to execute stage for higher util
This prevents coupling between C access in frontend & queue freeup in
backend.
2024-10-29 17:51:32 -07:00
Hansung Kim
37b8b6470b Bump vortex with race fix 2024-10-29 14:50:26 -07:00
Hansung Kim
216cfb0589 Bump vortex 2024-10-28 23:39:58 -07:00
Hansung Kim
752effdb21 tensor: Add FIFOFixer to smem tensor port
TODO: get area result for this
2024-10-28 23:38:56 -07:00
Hansung Kim
daabeb03ab tensor: Fix wrong addressGen that used bits not bytes 2024-10-28 22:27:34 -07:00
Hansung Kim
1ae1965580 tensor: Add IO and latching for smem address 2024-10-28 19:28:45 -07:00
Hansung Kim
c22fd20616 Bump vortex to 8cores 2024-10-27 19:47:52 -07:00
Hansung Kim
0e389dc362 Bump vortex 2024-10-27 18:49:42 -07:00
Hansung Kim
0ba61aabb6 tensor: Instantiate correct fake tcore module according to parameter
This has to align with what the verilog source actually uses.
2024-10-27 18:48:44 -07:00
Hansung Kim
13b9577723 Instantiate fake tensor modules outside of diplomacy
Adding them to the Diplomacy graph will cause to widen source widths
which can have area implications.

This gets rid of the need to do addResource() calls to the manually
generated Verilog files.  Their module parameters should be kept the
same as what's used in the parent Verilog module, however.
2024-10-25 23:02:25 -07:00
Hansung Kim
543eb2feb4 tensor: Support FP16 in TensorCoreDecoupled 2024-10-25 22:26:04 -07:00
Hansung Kim
eed821eda6 tensor: Add test for 8-dim fp16 DPU 2024-10-25 21:57:28 -07:00
Hansung Kim
46a57fdf9b tensor: Parameterize dimension in TensorDotProductUnit 2024-10-25 21:57:22 -07:00
Hansung Kim
51dfebb6a7 tensor: Support pipe = 1 in FillBuffer for higher throughput 2024-10-25 20:20:53 -07:00
Hansung Kim
d46a343239 tensor: Fix metadata of C req; fix dequeue / req gen timing 2024-10-25 19:13:42 -07:00
Hansung Kim
1a1a4a088d tensor: Fix access state transition to consider C req 2024-10-25 18:23:51 -07:00
Hansung Kim
991025e896 tensor: Fix C reg being dropped by checking space in respQueueC 2024-10-25 18:10:35 -07:00
Hansung Kim
81efecb3c8 tensor: Fix timing of fullCTag 2024-10-25 17:29:35 -07:00
Hansung Kim
43e064fe82 tensor: Add access logic for C from regfile 2024-10-25 15:22:52 -07:00
Hansung Kim
fc5b864b86 Bump vortex; addResource tensor regfile if 2024-10-24 20:35:14 -07:00
Hansung Kim
31fa440000 Bump vortex 2024-10-24 15:25:12 -07:00
Hansung Kim
ccfb467587 Bump vortex 2024-10-24 15:24:28 -07:00
Hansung Kim
988f0e3174 smem: Disable sanity check on partialData 2024-10-24 15:24:28 -07:00
Hansung Kim
f989bfccc2 Add tensorCoreDecoupled param to WithRadianceCores 2024-10-24 15:24:28 -07:00
Richard Yan
68e715e284 fix unaligned port 2024-10-24 13:42:45 -07:00
Richard Yan
9b8d16d184 Merge branch 'main' of https://github.com/ucb-bar/radiance into main 2024-10-23 15:09:48 -07:00
Richard Yan
0a54018650 dual read port srams 2024-10-23 15:09:43 -07:00