Closed Bug 707738 Opened 13 years ago Closed 6 years ago

Investigate why using CVTTSS2SI fast/slow path is not faster than integer_d(f2d(x))

Categories

(Tamarin Graveyard :: Baseline JIT (CodegenLIR), defect, P5)

Product:
Tamarin Graveyard
Component:
Baseline JIT (CodegenLIR)
Platform:
x86
macOS
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED WONTFIX
Milestone:
Q2 12 - Cyril

People

(Reporter: edwsmith, Unassigned)

References

Details

In fixing bug 707644, I first implemented a fast/slowpath that used CVTTSS2SI, only to find that promoting to double first, then using the non-inlined helper, was fastest. This bug is to deep-dive on why, and if possible, generate even faster code. test/performance/asmicro/float-toint-*.as are the tests I used to make the decision.
Blocks: float/float4
Priority: -- → P3
Target Milestone: --- → Q2 12 - Cyril
Priority: P3 → P5
Tamarin is a dead project now. Mass WONTFIX.
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → WONTFIX
Tamarin isn't maintained anymore. WONTFIX remaining bugs.
You need to log in before you can comment on or make changes to this bug.