-
```
Mozilla is using a SIMD-accelerated JPEG decorder in Fx5. We should try to
adapt this to VMX using what we will have learned from WebM. Medium as we're
still reasonably fast at this even in C.
…
-
```
* This is a startup crash or failure to start (Y/N): N
* This is a fault of JavaScript acceleration (Y/N): (Use the steps above to
find out. Do NOT report if it is not.) N
* What steps are neces…
-
```
Created in content/base/src modeled on the SSE2 version. On the G5 in reduced,
6.0 before the fix with the benchmark in
https://bugzilla.mozilla.org/show_bug.cgi?id=585978 yielded
1382ms
612ms
…
-
I still need to support a few PPC machines in my domain. For this, I keep old Universal-versions of applications around, which get distinguished from the newer versions available to Intel-Macs via the…
-
```
https://bugzilla.mozilla.org/show_bug.cgi?id=976446
It appears to use the existing MacroAssemblers (yay!), so we just need to adopt
the ARM LR usage, and probe for endian problems. At least it l…
-
```
Here are the changes necessary for building on Intel. I made an effort to make
the changes active only on Intel but I didn't try compiling the whole thing on
my poor old G4.
The font table cach…
-
```
https://bugzilla.mozilla.org/show_bug.cgi?id=692879
We can't compile Skia on 10.4. This might be nice, but the patch uses several
10.5-specific methods. If all else fails, we still have good old…
-
```
For 7400, 7450, G5,
-ftree-vectorize
For G5,
-mcpu=G5 -mno-powerpc64
and remove the stupid kludge we are using currently for 32-bit compiles.
These options are a little risky, so let's test them…
-
```
from Chris, on a 1.33GHz PowerBook,
Sunspider 1.0.2 (I ran these several times, the results were consistent)
==31==
Total: 6800.5ms +/- 1.1%
3d: 878.4ms +/- 2.9%
cube: 381.6ms +/- 6.7%
morph: 21…
-
```
Spun off from issue 191. After patching this out, overhead drops a bit, but
we're still spending ~40% of our time in semaphore lock on the affected sites.
Shark looks like this:
0.0% 95.…