667099 - SpiderMonkey is 2x faster than v8 on plb matmul benchmark

Status: RESOLVED INVALID

(Core :: JavaScript Engine, defect)

Description

[David Mandelin [:dmandelin]]

host-5-88:matmul dmandelin$ time d8 matmul_v1.js 
gc scavenge 2011238 0.754403
gc scavenge 2064450 0.774362
gc scavenge 737837 0.276758
gc scavenge 486846 0.182613
gc scavenge 170400 0.063916
gc scavenge 173625 0.065126
-9.3358333

real	0m0.076s
user	0m0.047s
sys	0m0.006s
host-5-88:matmul dmandelin$ time js -m -a  matmul_v1.js  
-9.3358333

real	0m0.141s
user	0m0.133s
sys	0m0.006s
host-5-88:matmul dmandelin$ time js -j  matmul_v1.js 
-9.3358333

real	0m0.139s
user	0m0.132s
sys	0m0.005s

host-5-88:matmul dmandelin$ time js  matmul_v1.js 
-9.3358333

real	0m0.144s
user	0m0.132s
sys	0m0.006s

As in bug 667096, jits seem to have little effect.

Comment 1

[David Mandelin [:dmandelin]]

Fixed my build:

host-5-88:opt32 dmandelin$ time shell/js -j ~/sources/plb/matmul/matmul_v1.js >
 /dev/null

real    0m0.025s
user    0m0.017s
sys     0m0.006s
host-5-88:opt32 dmandelin$ time shell/js -m -a ~/sources/plb/matmul/matmul_v1.j
s > /dev/null

real    0m0.032s
user    0m0.024s
sys     0m0.006s

Seems like a low priority bug. :-)

Comment 2

[Brendan Eich [:brendan]]

(In reply to comment #1)
> Seems like a low priority bug. :-)

Reopen and resummarize, then?

/be

Comment 3

[David Mandelin [:dmandelin]]

(In reply to comment #2)
> (In reply to comment #1)
> > Seems like a low priority bug. :-)
> 
> Reopen and resummarize, then?

I don't think there's anything to do here now. So I guess I should either leave it with this description and WONTFIX it, or else put back the original incorrect description and INVALID it. Any preference?