Everyone - please post OGR benches from 460 client here - for speeds page :)

[aberant]

Just thinking it might be pertiant seeing as you can benchmark with the new clients. Or, you could teach me what and .asp is and i could to do it ! hehe. Just a suggestion

 

[Mika]

Great idea! If someone could send me the OGR numbers for all the processor's listed here http://www.teamanandtech.com/speed.asp then I could have it updated in a matter of minutes!

Mika

 

[UmbrAC]

I got 6.4k nodes/s for my Athlon OGR long bench. Also, if you could change the RC5 for the athlon too.

 

[Mika]

Can somebody confirm the new Athlon RC5 multiplier is 3.4 before I change it?

Mika

 

[JonB]

Let's see. Since I love OGR so much.

For the 2.8009-460-CTR-00060723 client for Win32,

For Celeron 300a/450, using core #0 (GARSP 5.13), long bench

I get 2,557,661.16 nodes/sec

plus, its a pretty day outside and I'm not at work. Does it get much better than this????

 

[BurntKooshie]

Mika - i can confirm that its over 3.3, but at least in win98, under 3.4. I get 2.36mkeys/sec @ 700mhz. That's 3.37 kkeys/mhz. So with another OS, it might be able to squeak another .03 out of it...

 

[aberant]

MIKA - for OGR i get 736995 as average from 3 long OGR benches. Winchip C6 at 225 Soon to come - standard pentium, K6-2 and G$ benches

 

[UmbrAC]

I agree with the 3.4 for athlon rc5.

 

[Viztech]

OGR long benchmarks.
Averages from 4 runs

K6 233 1143099/233=4906 nodes/sec/mhz
K6-2 450 2082565/450=4627
Celeron 366@550 3046955/550=5539
Pentium Classic 166 602380/166=3628

viz

 

[crYnOid]

OGR long benchmark (average of 3 runs)
k6-3 450 2,176,156.75 nodes/sec (core #0)

 

[PeterN]

Did a run for a PIII 450@556.
The selected OGR core was #0.

OGR-long ==> 2,969,378.31 nodes/sec.
OGR-short ==> 2,946,536.44 nodes/sec.

Peter

 

[nukefarmer]

OGR benchmark on Celeron 433@468 (had to take it down a little because of the hot weather )
Avg : 2,575,154.08 nodes/sec

 

[aberant]

right - what we got so far...

Athlon - Confirmed
Celeron - Confirmed
Winchip
K6
K6-2
Pentium Classic
K6-3
PIII (katmai)

So we still need ...
386
486
Pentium MMX
PII
PPro
K5
All cyrix procs
All mac procs

 

[Ken g6]

Well, with BurntKooshie's Cyrix "upgraded" to heaven (or where? ) I guess I'll have to do the benchmarking honors.

594,663.36 Nodes/sec for a Cyrix 6x86 P150+@120MHz

 

[BurntKooshie]

Ken_G6... Hehe...nope, its serving as a dedicated crack rack Its vid card is SO bad, that it does 640 * 480 in 16 GREY colors That, and I don't have a mouse for it right now....nor a modem (or network card), nor an extra floppy, so give me another few days, and I'll get around to putting some of MY cyrix benchmarks up

 

[PeterN]

Tomorrow I'll do a run on a P166@200 MMX.
It's at work, so I'll have to wait.
The numbers will come eventually.

 

[Engine]

For my parents p233mmx:
[Jun 23 02:42:26 UTC] Benchmark for OGR core #0 (GARSP 5.13)
0.00:00:16.25 [959,500.64 nodes/sec]

 

[nukefarmer]

Also checked my p200-mmx here at work, this one does about 820,954.87 nodes/sec.

 

[SSP]

Athlon 600 @ 650:
[Jun 23 07:37:40 UTC] Benchmark for OGR core #0 (GARSP 5.13)
0.00:00:16.12 [4,184,365.13 nodes/sec]

I'll add the 486 results later.

 

[PeterN]

Here are the numbers for a P166@200 MMX:
OGR core #0
Long ==> 824,979.65 n/s
short ==> 799,082,39 n/s

 

[aberant]

i was thinking last night, should the pPro not be on a line of its own ? Just that the ppro didnt have mmx, which surely would slow it down at rc5, no ? just a thought

 

[Ken g6]

^^^

Actually, Aberant, the RC5 client doesn't use MMX - Osiris knows why. I guess there is no MMX rotl instruction or something.

Any more benchmarks?

 

[TheyCallMeSAK]

Alright, PIII 700 (no o/c)

[Jun 24 21:45:07 UTC] Benchmark for RC5 core #2 (RG class 6)
0.00:00:16.03 [1,961,925.90 keys/sec]
[Jun 24 21:45:26 UTC] Benchmark for OGR core #0 (GARSP 5.13)
0.00:00:16.05 [3,875,661.60 nodes/sec]
[Jun 24 21:45:41 UTC] Benchmark for RC5 core #2 (RG class 6)
0.00:00:08.05 [1,952,912.83 keys/sec]
[Jun 24 21:45:52 UTC] Benchmark for OGR core #0 (GARSP 5.13)
0.00:00:08.07 [3,911,675.13 nodes/sec]

 

[amok]

PIII 620 (Katmai) Avg. of 4 runs with core #0: 3.38M nodes/s

That translates out to about 5451 nodes/s/MHz.

 

[aberant]

Ken yes it does its SSE/3Dnow that it doesnt use...

"2.7100.418 new: RC5 MMX cores are now in place. 50% speed
improvement on Pentium MMX processors."

taken from changes.txt