Blaise Tine
|
e770824d47
|
fixed afu cci write bug, fixed profile cache write miss bug, fixed bram byteenable inferance
|
2021-01-10 20:26:15 -08:00 |
|
Blaise Tine
|
8aea9cbe07
|
minor update
|
2021-01-06 21:39:15 -08:00 |
|
Blaise Tine
|
2b8435471a
|
speeding up simulation using dedicated full dpi-based FPU core
|
2021-01-06 18:44:06 -08:00 |
|
Blaise Tine
|
2d69ca5d67
|
scratchpad optimization for stack access using custom bank offset aligned to stack size
|
2021-01-02 16:00:00 -05:00 |
|
Blaise Tine
|
30d950ada2
|
vx_spawn_warps redesign using opencl's style scheduler
|
2021-01-01 14:13:48 -05:00 |
|
Blaise Tine
|
e4a00dd0d9
|
fixed loader script stack setup
|
2020-12-31 22:37:20 -05:00 |
|
Blaise Tine
|
d44144f72f
|
FPU float<->int conversion optimization
|
2020-12-29 15:37:45 -08:00 |
|
Blaise Tine
|
e83c4638a0
|
FPU area optimization sharing fmadd hard block
|
2020-12-27 17:31:10 -08:00 |
|
Blaise Tine
|
25df233005
|
Adding Altera Stratix 10 support
|
2020-12-27 10:44:57 -08:00 |
|
Blaise Tine
|
b2b8f190dd
|
minor update
|
2020-12-26 14:47:41 -08:00 |
|
Blaise Tine
|
33c431ed44
|
multiplier unit optimization - using fifo for metadata, shift register optimization
|
2020-12-26 11:23:21 -08:00 |
|
Blaise Tine
|
b459192dec
|
critical path optimization - fpga fmax @4c = ~212 mhz
|
2020-12-26 03:28:32 -08:00 |
|
Blaise Tine
|
703a861fe9
|
added support for write-through cache, removed cache snooping support
|
2020-12-23 23:51:02 -08:00 |
|
Blaise Tine
|
d956e268b9
|
adding new performance counters (banks utilization and DRAM bus utilization)
|
2020-12-22 12:33:45 -08:00 |
|
Blaise Tine
|
4b7d871d62
|
allowing partial cache request submissions, io bus support broken
|
2020-12-21 03:53:13 -08:00 |
|
Blaise Tine
|
4bbd7bf408
|
performance refactoring - rebalanced stream buffers accross the device to enforce output buffering rule at compoments boudaries, finally resolved block ram R/W collusion discrepencies,
|
2020-12-19 02:45:06 -08:00 |
|
Blaise Tine
|
d5438fd591
|
merging perf counters
|
2020-12-08 21:02:39 -08:00 |
|
Xandy Liu
|
1595ff08e2
|
PERF pipeline stalls and cache
|
2020-12-08 01:14:41 -05:00 |
|
Blaise Tine
|
268ad15098
|
minor update
|
2020-12-06 22:55:17 -08:00 |
|
Blaise Tine
|
d68b32cd60
|
minor update
|
2020-12-06 22:40:27 -08:00 |
|
Blaise Tine
|
1332970636
|
refactoring cores clustering
|
2020-12-06 14:42:12 -08:00 |
|
Blaise Tine
|
b2652527bb
|
data/dram bus refactoring
|
2020-12-06 03:37:22 -08:00 |
|
Blaise Tine
|
478d971389
|
minor update
|
2020-12-03 16:21:20 -08:00 |
|
Blaise Tine
|
0a8f41964d
|
minor update
|
2020-12-03 08:47:03 -08:00 |
|
Blaise Tine
|
b7a724410b
|
update DRAM simulation - reduce the latency of duplicate requests (simulate DRAM cache)
|
2020-12-03 07:30:19 -08:00 |
|
Blaise Tine
|
f68af3bb84
|
using mshr pending request size
|
2020-12-01 00:54:25 -08:00 |
|
Blaise Tine
|
def6a35693
|
shared memory optimization
|
2020-11-29 15:04:31 -08:00 |
|
Blaise Tine
|
b85391389b
|
rename MSRQ to MSHR
|
2020-11-28 17:32:00 -05:00 |
|
Blaise Tine
|
00d7473268
|
build warnings clean
|
2020-11-28 14:59:13 -05:00 |
|
Blaise Tine
|
7ae770f4eb
|
config update
|
2020-11-21 12:27:42 -08:00 |
|
Blaise Tine
|
1795980a52
|
L2 and L1 using different block size support, RTLsim fixes, dram_rsp_ready optimization
|
2020-11-21 09:47:56 -08:00 |
|
Blaise Tine
|
4d94d33e8d
|
constant integration updates
|
2020-11-15 09:12:48 -08:00 |
|
Blaise Tine
|
725322807e
|
fixed DRAM response backpressure inside Cache
|
2020-11-10 05:24:57 -08:00 |
|
Blaise Tine
|
10505caae1
|
refactoring all arbiters with buffering for request count > 2, optimized the cache core response module in critical path when running as L2
|
2020-11-08 01:31:46 -08:00 |
|
Blaise Tine
|
ba81d76e02
|
cache refactoring - phase 2
|
2020-11-03 04:51:40 -08:00 |
|
Blaise Tine
|
5be1d85648
|
cache refactoring (fixed redundant fill requests, merged fill and writeback queues), optimized priority encoder, fixed crs cycles count
|
2020-11-02 01:50:12 -08:00 |
|
Blaise Tine
|
9a9f7955f0
|
basic test timing + scope tracing ccip
|
2020-10-27 17:04:04 -04:00 |
|
Blaise Tine
|
7529f72c5d
|
fixed OPAE crash, added custom bram module to controll rw collision, dogfood testcase argurment, optimzed buffered fifo, quartus build optimization flags
|
2020-10-20 05:32:55 -07:00 |
|
Blaise Tine
|
f6f95e0c46
|
mutiple fixes: parallel printf, fixed cycle in cache, opencl refactored vecadd and sgemm, regen opencl kernels with hard-float, fixed vortex io bus interface, fixed dpi floats APi to support multicore mode, make vlsim multicore default, make rtlsim multi-core default, removed POCL binaries from repository, updated Makefiles to use external POCL
|
2020-09-19 14:45:42 -04:00 |
|
Blaise Tine
|
36ec603d17
|
fpga fixes: warp scheduler, fnmadd, fdiv, fsqrt
|
2020-09-08 07:05:26 -07:00 |
|
Blaise Tine
|
49b86c4b2a
|
SCOPE update
|
2020-09-05 10:52:59 -07:00 |
|
Blaise Tine
|
42e3b6c45d
|
fixed lmp_mult parameters, ram init filepath
|
2020-09-04 07:51:46 -07:00 |
|
Blaise Tine
|
31ffbe0d6a
|
clean up 'stage_1_cycles' from cache
|
2020-09-01 03:39:03 -07:00 |
|
Blaise Tine
|
0a45a8beb3
|
minor update
|
2020-09-01 00:56:10 -07:00 |
|
Blaise Tine
|
4e8b9fb296
|
FPU SVDPI support complete
|
2020-09-01 00:59:37 -04:00 |
|
Blaise Tine
|
c1df08843c
|
minor update
|
2020-08-31 09:34:19 -04:00 |
|
Blaise Tine
|
df711986bc
|
FPU DPI fallback
|
2020-08-31 09:19:55 -04:00 |
|
Blaise Tine
|
fde3f46798
|
ibuffer optimization
|
2020-08-26 04:44:36 -07:00 |
|
Blaise Tine
|
1c9445745f
|
fp_noncomp fixes
|
2020-08-23 16:53:28 -07:00 |
|
Blaise Tine
|
0b355f228e
|
ibuffer addition
|
2020-08-22 00:22:04 -07:00 |
|