 Constantin Fürst
|
d1cc3e3b0c
|
modification to qdp benchmark, returns to per-chunk barrier wait, uses userspace semaphore for one-way barrier from scan_b to aggr_j as scan_b should submit asap but aggr_j should wait on submission from scan_b, contains TODO for modifying code to support chunkcount not divisible by 2
|
11 months ago |
|
Constantin Fürst
|
a963406f7c
|
move mode selection to Configuration.hpp, adapt the CopyMethodPolicy-Function to return only src_node for task sizes under 16MiB which is now required to not cause high submission count which slows down small copies
|
11 months ago |
|
Constantin Fürst
|
ef805244ac
|
use 4gib as size and again 1 aggrj thread for qdp bench
|
11 months ago |
|
Constantin Fürst
|
81527fdb6b
|
commit current vampir config
|
11 months ago |
|
Constantin Fürst
|
69a3d2cef4
|
experimental implementation for tc-scanb > tc-aggrj, the second
|
11 months ago |
|
Constantin Fürst
|
07fba8a5f0
|
experimental implementation for tc-scanb > tc-aggrj
|
11 months ago |
|
Constantin Fürst
|
d4122ba25a
|
add updated config for prefetch from vampir
|
11 months ago |
|
Constantin Fürst
|
e4a0030049
|
fix prefetching subchunk indexing and adapt the weak access flag for join
|
11 months ago |
|
Constantin Fürst
|
972440d19f
|
repair flags implementation
|
11 months ago |
|
Constantin Fürst
|
b3607329a6
|
add a flags-concept to cacher, add the option to select whether to handle pagefaults or not
|
11 months ago |
|
Constantin Fürst
|
bb1d20924a
|
fix index clash for thread-and-group unique indexing
|
11 months ago |
|
Constantin Fürst
|
0eca180e53
|
fix destination indexing in aggrj for happly
|
11 months ago |
|
Constantin Fürst
|
5e8f3e05e3
|
fix chunk indexing in scanb and refactor result calculation
|
11 months ago |
|
Constantin Fürst
|
c2b9e6656d
|
fix chunk selection in scanb, use the dataptr in aggrj complex mode, export some functions to src/utils/BenchmarkHelpers.cpp
|
11 months ago |
|
Constantin Fürst
|
845e812ca7
|
set the correct sum check which was inverted by querry type
|
11 months ago |
|
Constantin Fürst
|
abcb9a4b2e
|
extend modestring to contain query type
|
11 months ago |
|
Constantin Fürst
|
e4ed4ac5b9
|
correct and minimize subchunking implementation which now is only allowed in scanb
|
11 months ago |
|
Constantin Fürst
|
50560606a3
|
add complex query as benchmarking option and evaluate results
|
11 months ago |
|
Constantin Fürst
|
3c1606da51
|
init datab correctly as well to fix the benchmark
|
11 months ago |
|
Constantin Fürst
|
10a791dea1
|
remove the experimental code branches that turned out not to yield any benefit (sched-yield has too high delay and with the new load balancer, subchunking for aggrj is also not needed anymore)
|
11 months ago |
|
Constantin Fürst
|
881047068c
|
rerun benchmarks for dram baseline and hbm peak
|
11 months ago |
|
Constantin Fürst
|
a72a26dbee
|
remove cout/cerr output from cache and benchmark to not falsify results
|
11 months ago |
|
Constantin Fürst
|
0856d58855
|
properly drop barrier when using iterrative aggregation
|
11 months ago |
|
Constantin Fürst
|
5c08313830
|
properly set barrier when using iterrative aggregation
|
11 months ago |
|
Constantin Fürst
|
178d45fafa
|
use weak wait, add options to tweak for caching mode
|
11 months ago |
|
Constantin Fürst
|
34f7aca50a
|
correct bad thread timing storage size set which should have been 1 from the start and not 0
|
11 months ago |
|
Constantin Fürst
|
100774f495
|
remove the step-by-step barrier sync and let scana and scanb run to completion before starting with aggrj
|
11 months ago |
|
Constantin Fürst
|
52aaab3c09
|
prevent illegal instruction exception when no measurements have been conducted
|
11 months ago |
|
Constantin Fürst
|
391e6ca273
|
use proper timing indices for documenting thread runtime
|
11 months ago |
|
Constantin Fürst
|
e429e8fd40
|
adapt barrier waiting points, add timings to thread execution
|
11 months ago |
|
Constantin Fürst
|
2fe9aa90e3
|
latest changes to benchmarker
|
11 months ago |
|
Constantin Fürst
|
542a815b4f
|
adapt to efficient wait in cache
|
11 months ago |
|
Constantin Fürst
|
5455f51f61
|
properly handle more scanb than aggrj threads
|
11 months ago |
|
Constantin Fürst
|
b49bc6c173
|
push changes to benchmark
|
11 months ago |
|
Constantin Fürst
|
3dee13b3bc
|
transfer commit from crobat
|
11 months ago |
|
Constantin Fürst
|
4a6529b111
|
revamp chunk calculation
|
11 months ago |
|
Constantin Fürst
|
6502f95bb2
|
correct mode naming of the execution modes
|
11 months ago |
|
Constantin Fürst
|
2f9d059252
|
add prelimianry results, modify the launch script and provide macro-based selection of parameters for the three modes {dram,hbm,prefetch}
|
11 months ago |
|
Constantin Fürst
|
1820166e6f
|
slight refactoring of qdp benchmark and add mandatory wait to cachedata in aggrj
|
11 months ago |
|
Constantin Fürst
|
52132522a3
|
fix scanb not working with less than aggrj threads
|
11 months ago |
|
Constantin Fürst
|
e34b4df7e6
|
fix wrong linebreak in result writer
|
11 months ago |
|
Constantin Fürst
|
f1ac2a07b2
|
write result to ofile in qdp bench
|
11 months ago |
|
Constantin Fürst
|
457a3b520a
|
add barriers to the qdp benchmark
|
11 months ago |
|
Constantin Fürst
|
aa0867aa3a
|
reworking the qdp benchmark
|
11 months ago |
|
Constantin Fürst
|
ab217cb080
|
remove bad opt mode from last commit and instead try to improve default prefetching
|
11 months ago |
|
Constantin Fürst
|
2e0f637363
|
add optimal dsa caching mode
|
11 months ago |
|
Constantin Fürst
|
f16d67f67e
|
wait per iteration for caching mode
|
11 months ago |
|
Constantin Fürst
|
5c896dbf04
|
publish best run config for dsa-prefetch
|
11 months ago |
|
Constantin Fürst
|
0f843d9282
|
potential fix for lifetime of cache data take two
|
11 months ago |
|
Constantin Fürst
|
74f659e5a4
|
potential fix for lifetime of cache data
|
11 months ago |
