yuzu-emu/FasTC
1
0
Fork 0
mirror of https://github.com/yuzu-emu/FasTC.git synced 2024-11-24 02:35:38 +01:00
Code Issues Packages Projects Releases Wiki Activity

Another bug fix.

Browse Source 
With the old code, it was possible that we skipped a compression with unlucky
preemption of our threads. I'm not exactly sure why, but that caused deadlock
(livelock?) in some very unfortunate circumstances. This new algorithm should
work regardless of how many threads execute at once and should also prevent
textures in the compression job list from being skipped. This algorithm seems
to be an improvement on low-core count machines (around 4 cores), but it is
slower on high-core count machines (40 cores or more)...
This commit is contained in:
Pavel Krajcevski 2013-03-11 16:20:52 -04:00
parent 9c48aaa7f2
commit 6f6ca2d867
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
BPTCEncoder/src/BC7Compressor.cpp
View File

@ -1555,7 +1555,7 @@ namespace BC7C
// Help finish whatever texture we're compressing before we start again on my work...
uint32 blockIdx;
while((blockIdx = FetchAndAdd(&(cjl.m_CurrentBlockIndex))) < nBlocks && jobIdx == cjl.m_CurrentJobIndex) {
while((blockIdx = FetchAndAdd(&(cjl.m_CurrentBlockIndex))) < nBlocks && *(cjl.GetFinishedFlag(jobIdx)) == 0) {
unsigned char *out = cj->outBuf + (16 * blockIdx);
const unsigned char *in = cj->inBuf + (64 * blockIdx);
@ -1563,8 +1563,8 @@ namespace BC7C
}
if(TestAndSet(cjl.GetFinishedFlag(jobIdx)) == 0) {
cjl.m_CurrentJobIndex++;
cjl.m_CurrentBlockIndex = 0;
cjl.m_CurrentJobIndex++;
}
// Wait until this texture finishes.