@RKSimon this good now?
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Today
Yesterday
ping 4 @nikic
Rebase
Rebase
Rebase
Rebase
In D145219#4230166, @RKSimon wrote:It'd be useful to add tests to see if the relevant ternlog patterns get correctly canonicalized to the useful patterns listed here: http://www.0x80.pl/articles/avx512-ternary-functions.html
In particular - ensuring we fold ternlog -> bitselect -> regular select (if driven by a compare / sign-extended bool vector), and maybe the one-bit-set pattern are of particular interest?
Remove unnecessary NTests loop
Add some known patterns
Tue, Mar 28
Rebase
Split of (shl Z, X), (shl Y, X) case. Remove the Shift{Y|Z|X} flags
Add unit tests
Rebase
Sun, Mar 26
In D145280#4182474, @nikic wrote:Is there any real-world motivation for these xor transforms (and also the signed predicate variants in this patch stack)? If not, I'm not sure this is worthwhile.
In D146350#4204295, @nikic wrote:This is not generally legal for div/rem, see https://alive2.llvm.org/ce/z/2iiR26 using one of your tests. This is only legal if the operation is speculatable.
Wed, Mar 22
In D142666#4213821, @RKSimon wrote:In D142666#4213794, @goldstein.w.n wrote:In D142666#4212405, @RKSimon wrote:@goldstein.w.n reverse-ping - whats the plan for this patch?
@RKSimon does this show up in any codegen that went throw the middle-end?
There isn't that much in DAG that generates ABS nodes - foldAndOrOfSETCC is the closest to this I suppose?
In D142666#4212405, @RKSimon wrote:@goldstein.w.n reverse-ping - whats the plan for this patch?
In D142666#4212405, @RKSimon wrote:@goldstein.w.n reverse-ping - whats the plan for this patch?
Mon, Mar 20
Sat, Mar 18
Rebase
Rebase
Rebase
Propegate nuw flag
Add flag propegation test
Rebase
Rebase
Rebase
Rebase
Put tests in canonical commuted form. Add multiuse tests
Fri, Mar 17
Thu, Mar 16
In D141653#4198903, @RKSimon wrote:@goldstein.w.n What is happening with this patch? After they leave my "Ready to Review" list I tend to lose track......
The new IR with and removes a use of the input variable, which is better for analysis.
"and" -> And (otherwise it looks like the conjugation, not the operator.
Wed, Mar 15
ping.
LGTM.
Mon, Mar 13
Rebase
Rebase
Fix nits
Rebase
Wed, Mar 8
Tue, Mar 7
Rebase
Add common check for AVX2
Rebase
Rebase
Nit + ICX
In D144570#4174550, @RKSimon wrote:Add ICX test runs to tuning-shuffle-unpckpd.ll so we have test coverage ?