wu-arch/vortex
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Commit Graph

  • 2743d32bd2 tensor: Handle wid queue backpressure in dpu Hansung Kim 2024-05-30 15:25:00 -07:00
  • 2e2decc8b6 Shrink size of D_half latch Hansung Kim 2024-05-30 12:46:45 -07:00
  • 73a2f5781e Do two-cycle compute with 1 FEDP per lane Hansung Kim 2024-05-29 22:01:03 -07:00
  • 35273b3d74 Set correct dpu hmma latency Hansung Kim 2024-05-29 17:14:54 -07:00
  • 5ed6041e33 tensor: Properly stall dpu upon commit backpressure Hansung Kim 2024-05-29 17:05:12 -07:00
  • f5a9ca5bf3 tensor: Enqueue both insts in pair to issue queue Hansung Kim 2024-05-29 14:47:25 -07:00
  • e9df173745 tensor: Use chisel-generated dpu module Hansung Kim 2024-05-29 13:34:25 -07:00
  • c03a5b070c tensor: Issue queue for dpu to improve utilization Hansung Kim 2024-05-27 18:24:24 -07:00
  • 28f6cd59b5 tensor: Improve commit efficiency by decoupling dpu with fifo Hansung Kim 2024-05-26 21:59:25 -07:00
  • 864265bda5 tensor: Fix consecutive commits to write to same warp Hansung Kim 2024-05-25 20:04:31 -07:00
  • 5a95eba1f5 tensor: Clear c_*_tile before compute Hansung Kim 2024-05-25 19:54:03 -07:00
  • 8775458a8f Stage half-operands per warp Hansung Kim 2024-05-25 19:08:17 -07:00
  • 45d86b26a2 tensor: Add counter for dpu operations Hansung Kim 2024-05-16 22:15:01 -07:00
  • 5034d8d14b tensor: Add buffer to hide 2cyc commit latency Hansung Kim 2024-05-16 20:07:30 -07:00
  • 317695a8d0 Add perf counters on LSU resp valid tmasks Hansung Kim 2024-05-16 15:32:46 -07:00
  • 89e7d65926 tensor: Add ready signal to enforce 1 warp occupancy Hansung Kim 2024-05-16 12:49:15 -07:00
  • 1a1094b2bb tensor: Add dispatch unit to narrow to BLOCK_SIZE=1 Hansung Kim 2024-05-15 15:34:26 -07:00
  • 9f9ec10960 tensor: Enable scaling NUM_THREADS by octets Hansung Kim 2024-05-08 11:26:09 -07:00
  • d624b3e50a store fencing, large smem, fix tensor core for firesim Richard Yan 2024-05-15 21:45:48 -07:00
  • 0dd5335851 fix merge error once again Richard Yan 2024-05-08 11:31:43 -07:00
  • 16dfae7d3f Merge branch 'rtl' of https://github.com/hansungk/vortex-private into rtl Richard Yan 2024-05-08 11:28:39 -07:00
  • 629279977e fix merge error Richard Yan 2024-05-08 11:28:36 -07:00
  • be748b109a Fix faulty merge on syn-only flags Hansung Kim 2024-05-07 18:37:25 -07:00
  • f71e705d53 Revert to old LSUQ_SIZE Hansung Kim 2024-05-07 16:23:32 -07:00
  • 4aad161739 Merge branch 'rtl' of https://github.com/hansungk/vortex-private into rtl Richard Yan 2024-05-07 14:00:31 -07:00
  • 37616f3334 firesim modifications Richard Yan 2024-05-07 13:59:25 -07:00
  • c9a3eaad79 accelerator cisc Richard Yan 2024-05-07 13:58:32 -07:00
  • 14d1552f08 potential deadlock Richard Yan 2024-05-07 13:56:51 -07:00
  • 1e5dff52c1 shrink queue sizes Richard Yan 2024-05-07 13:54:23 -07:00
  • 868bbdb15e tensor: more doc Hansung Kim 2024-05-07 13:54:10 -07:00
  • b70df8cbc9 proper srams Richard Yan 2024-05-07 13:52:07 -07:00
  • 9c1d797250 tensor: add missing } Hansung Kim 2024-05-05 18:36:15 -07:00
  • fb626ee21c tensor: doc Hansung Kim 2024-05-05 18:35:52 -07:00
  • 9ea291eea2 Merge remote-tracking branch 'origin/tensor_core' into rtl Hansung Kim 2024-05-05 17:03:57 -07:00
  • 5bd25985c6 i kinda forgot most of changes joshua 2024-05-04 23:01:47 -07:00
  • 1c7acab160 tensor: Fix lint errors Hansung Kim 2024-05-03 15:43:02 -07:00
  • 5a0ee98a61 Remove duplicate port connection Hansung Kim 2024-05-03 15:07:24 -07:00
  • bc45c40231 tensor: Rename half.hpp -> half.h Hansung Kim 2024-05-01 22:01:34 -07:00
  • c4b94e4f2c Wrap hardcoded configs with SYNTHESIS Hansung Kim 2024-05-01 21:57:05 -07:00
  • c4d71bc3d6 tensor: Fix multiple driver error on VCS Hansung Kim 2024-05-01 21:36:12 -07:00
  • 7fc5b6a374 tensor: Fix elaboration error on VCS Hansung Kim 2024-05-01 21:35:53 -07:00
  • 675e8ea130 Merge branch 'tensor_core' into rtl Hansung Kim 2024-05-01 16:18:14 -07:00
  • 9a688a05b1 Add (unconnected) FPU perf counters Hansung Kim 2024-04-29 15:20:55 -07:00
  • 100fbbc048 Increase FPUQ_SIZE Hansung Kim 2024-04-29 15:19:48 -07:00
  • 85213d2876 synthesizable design Richard Yan 2024-04-17 18:05:51 -07:00
  • 17fd29c114 Merge branch 'rtl' of https://github.com/hansungk/vortex-private into rtl Richard Yan 2024-04-16 23:03:04 -07:00
  • 8de5470da4 round robin warp scheduling Richard Yan 2024-04-16 23:03:00 -07:00
  • 217bc189da ifdef-guard VX_operand* to enable including both in Chisel Hansung Kim 2024-04-15 22:06:47 -07:00
  • 4752b86858 Limit NUM_SFU_LANES to 4 Hansung Kim 2024-04-15 21:47:54 -07:00
  • 978b1fe2d0 Add operands stage with duplicated RF for rs1/2/3 Hansung Kim 2024-04-15 16:45:59 -07:00
  • 87b966a5fa Add perf counter for stall by any operand hazard Hansung Kim 2024-04-15 01:01:26 -07:00
  • 7ae54bd280 Remove unused IO in core_wrapper Hansung Kim 2024-04-13 17:13:25 -07:00
  • d3e0f18fd5 Merge branch 'rtl' of https://github.com/hansungk/vortex-private into rtl Richard Yan 2024-04-09 19:55:11 -07:00
  • 41a79a03a4 parametrize memory interface in core wrapper and update config.vh Richard Yan 2024-04-09 19:55:06 -07:00
  • 6c632200d5 Divide by per-breakdown cycle for avg stall cycles Hansung Kim 2024-04-03 15:29:51 -07:00
  • 62c7d1f4cf Report any fire cycles from scoreboard as well Hansung Kim 2024-03-29 12:23:15 -07:00
  • 50263a5f7d Rename sched_barrier_stalls -> perf_sched_barrier_idles Hansung Kim 2024-03-28 22:45:12 -07:00
  • d8f9359fae test case update joshua 2024-03-28 13:04:02 -07:00
  • 08d7721e11 annoying swizzling problems joshua 2024-03-28 03:00:15 -07:00
  • e16584ddd9 bleh still not work joshua 2024-03-27 00:26:04 -07:00
  • dd90736382 Reformat perfcount report Hansung Kim 2024-03-23 01:07:34 -07:00
  • 3e6a9a6104 Expose scoreboard fires to perf interface Hansung Kim 2024-03-23 01:06:40 -07:00
  • d99295793c Periodically report perf counter; reformat operand/FU stalls Hansung Kim 2024-03-23 00:02:02 -07:00
  • 83e151a189 Add valid / fire / cycles-issued perf counters to dispatch Hansung Kim 2024-03-23 00:01:15 -07:00
  • 573be030c8 Add issue-stall-by-operand-hazard perf counters Hansung Kim 2024-03-23 00:00:08 -07:00
  • dda67da84c Add issue-stall-by-unit-busy perf counters Hansung Kim 2024-03-21 18:11:12 -07:00
  • 3718a57937 Docs Hansung Kim 2024-03-21 15:44:50 -07:00
  • b254281295 initial tcore impl joshua 2024-03-21 01:29:38 -07:00
  • 9438862389 Add perf counter for barrier schedule stalls Hansung Kim 2024-03-20 15:29:28 -07:00
  • f9b4509936 initial tensor core joshua 2024-03-20 02:46:00 -07:00
  • 978dd3bdfe seemingly working fp32 implementation joshua 2024-03-19 17:56:59 -07:00
  • 7014ae24da Prettier perf count reports Hansung Kim 2024-03-19 15:25:46 -07:00
  • b25deb8a2e Fix assignment for perf counters Hansung Kim 2024-03-19 14:06:44 -07:00
  • df4b21507e Customize global barrier response logic for clusters Hansung Kim 2024-03-18 14:30:32 -07:00
  • 2525df9c5f Use GBAR_CLUSTER_ENABLE to guard cluster-specific modification Hansung Kim 2024-03-17 18:24:04 -07:00
  • 7f8abe99ff Fix wrong multicore parametrization in wrapper Hansung Kim 2024-03-17 18:23:09 -07:00
  • 40e2888733 Connect core gbar signals in wrapper Hansung Kim 2024-03-17 14:09:43 -07:00
  • 28f54bde7f Merge remote-tracking branch 'sungwoong/master' into rtl Hansung Kim 2024-03-14 09:15:59 -07:00
  • bd67ff3439 Fix creating bogus mem reqs when commit is stalled Hansung Kim 2024-03-13 20:43:27 -07:00
  • 8317a3fbe5 Fix fence by disallowing x-initialization instead of all-0 mask Hansung Kim 2024-03-07 17:39:18 -08:00
  • 010c4675ce Fix undeclared mem_perf_if Hansung Kim 2024-03-07 15:00:43 -08:00
  • b63333a4ec Merge remote-tracking branch 'upstream/master' into vortex2 Hansung Kim 2024-03-07 14:45:48 -08:00
  • beb3dce46d integer reduction unit joshua 2024-03-06 01:39:17 -08:00
  • e7b0a149c7 Print TAG_ONLY_WIDTH of req_tag in trace Hansung Kim 2024-03-04 21:10:59 -08:00
  • 3c2a266d37 second pass Sungwoong Ha 2024-03-01 21:27:26 -08:00
  • a9709edae2 first pass Sungwoong Ha 2024-03-01 21:05:52 -08:00
  • be7d87c82d temp Sungwoong Ha 2024-02-22 16:31:42 -08:00
  • 5f2b10b8a6 minor update Blaise Tine 2024-02-09 21:20:23 -08:00
  • 3fee1a6193 minor update Blaise Tine 2024-02-09 20:34:44 -08:00
  • ae7b01405c CI minor update Blaise Tine 2024-02-08 14:10:00 -08:00
  • be0db6e1a5 minor update Blaise Tine 2024-02-04 20:32:05 -08:00
  • 50028c1a33 Merge remote-tracking branch 'origin' into develop Blaise Tine 2024-02-04 20:19:30 -08:00
  • 8d4b6c804f minor update Blaise Tine 2024-02-04 20:17:12 -08:00
  • 6f7a389a1f arbiters unlock refactoring Blaise Tine 2024-02-04 20:16:18 -08:00
  • fe15647f98 minor update Blaise Tine 2024-02-04 02:11:53 -08:00
  • b0b7cd2b1e minor updates Blaise Tine 2024-02-03 19:09:53 -08:00
  • eb63767051 Don't hardcode SIMULATION Hansung Kim 2024-02-01 23:58:06 -08:00
  • 48558982f7 Merge remote-tracking branch 'upstream/master' into vortex2 Hansung Kim 2024-02-01 23:35:58 -08:00
  • a15f4fd483 [BUGFIX] Set mem_req_mask to 0 for fence Hansung Kim 2024-02-01 21:16:54 -08:00
  • f9cd8be19e minor update Blaise Tine 2024-01-31 13:35:43 -08:00