"Just a few more"

Richard M
Richard M
Joined: 11 Nov 04
Posts: 84
Credit: 587618481
RAC: 1624360
Topic 189683

Wish I could get a few more WU's like this one.

:)
Richard

arashi
arashi
Joined: 12 Jun 05
Posts: 37
Credit: 41761784
RAC: 0

"Just a few more"

Quote:

Wish I could get a few more WU's like this one.

:)
Richard

Yeah, its quite remarkable what the Finns squeeze out of a Nokia smartphone.

Blank Reg
Blank Reg
Joined: 18 Jan 05
Posts: 228
Credit: 40599
RAC: 0

Some thing is wrong cause a

Some thing is wrong cause a p4 3.0 is not that slow it should complete one in about 40000 ht or without ht in about 25000.

Hefto99
Hefto99
Joined: 11 Feb 05
Posts: 9
Credit: 461131800
RAC: 1570986

Yeah, and benchmarks for host

Yeah, and benchmarks for host id 217081 are also very interesting....
13368.49 million ops/sec for integer speed ...

drweaser
drweaser
Joined: 11 Nov 04
Posts: 2
Credit: 878769
RAC: 0

Hey wait a second, I think

Hey wait a second, I think that's my host! (well one of them)

Impressive isn't it :)

It is a dual processor 2.0 G5 mac, running the optimized client from team macnn.....even I thought the numbers were a little out of whack. Too bad I couldn't get the 400+ it claimed for a unit.....

Paul D. Buck
Paul D. Buck
Joined: 17 Jan 05
Posts: 754
Credit: 5385205
RAC: 0

If they used as "smart"

If they used as "smart" enough optimizer it will remove large portions of the benchmark code as it is "make-work" and if recognized is simply removed. There was a similar issue with the 4.19 vs. later where the optimizer's removal of code meant higher benchmark scores.

This gives rise to the interest to have the highest benchmark which results in higher than normal claims. Though averaging does bring those down some, it does lead to a slight upward "bias" ...

Anyway ... just food for thought ... I would go on but not sure you all care ... :)

drweaser
drweaser
Joined: 11 Nov 04
Posts: 2
Credit: 878769
RAC: 0

Actually, it is optimized to

Actually, it is optimized to use the altivec part of the G4+ cpus. So, I imagine it is still running the benchmarks, but doing so on the vector portion of the processor and getting a pretty good boost out of it. I also remember that whole benchmark compiling out thing and beleive they set up the code so that could no longer occur using normal tools (gcc, which Apple currently uses). The use of altivec enabled benchmarks is actually quite reasonable since the beta einstein client also uses altivec optimizations (and I am running that also).

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.