Hashmaps are super expensive and there's no reason not to use a vector - we
already have compact block IDs so we don't benefit from the sparseness. Huge win
for very little effort.
Spotted when profiling FEX. CondJump() in the JIT was almost 4% of our time (?!)
and all because of map slowness. Easy fix.
Difference at 95.0% confidence
-0.0196494 +/- 0.00194956
-3.92827% +/- 0.389753%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
..and be consistent about signedness.
clang doesn't seem able to do this itself.
Difference at 95.0% confidence, n=100
-0.272326% +/- 0.242353%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
now obsolete!
Results for the whole series are excellent:
Difference at 95.0% confidence
-3.97603% +/- 0.254656%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Augment IREmitter to inline constants as we generate code, rather than
needing a later clean up pass. This replaces the last function of ConstProp.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Now executable page tracking is implemented, this limit is technically
unnecessary and effectively never hit in normal code. Keep a reasonable
limit however to avoid accidentally inlining tail calls and exploring
dead branches in obfuscated code.
In b8dd5d95b ("OpcodeDispatcher: optimize X87FTWTag"), we optimized
X87FTWTag using an efficient Morton interleave operation. Here, we do the
inverse, optimizing SetX87FTW using an efficient Morton deinterleave
operation.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
I don't think these are worth it, and also currently they don't trigger ever.
n=100:
Difference at 95.0% confidence
-0.00245961 +/- 0.00139573
-0.524468% +/- 0.297615%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
This is expensive and only needed for spilling, so only do it for spilling. This
complicates the RA a bit but speeds us up on average since most blocks
don't spill. Total results of this change (including the prep commits that
slowed things down temporarily):
Difference at 95.0% confidence
-1.71952% +/- 0.455996%
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
if we spill for SRA, we don't need/want to execute this code path. this will be
load bearing by the end of this series.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Only the last prefix byte is retained when multiple are set. We were
accidentally generating a mask.
Additionally with 64-bit code, the legacy segment prefixes don't
overwrite if FS or GS have been set. So no weird behaviour where FS/GS
is set, a legacy prefix is used for padding, and then it "ignores" a bad
prefix by ignoring only the latest one.
Ryan Houdek