VOGONS


First post, by andrea

User metadata
Rank Member
Rank
Member

Hello to all,
I remember reading on how the Copppermine Celeron, even at high clocks, had a considerable performance deficit compared to a P3.
These past few days I've been playing with one, a quite lucky 566 that manages a 100% overclock with stock vcore.

LLSmoe7.png

My "problem" is admittedly a weird one: Is this thing supposed to perform this good?

These Sandra (2002) results seem really weird to me:
v4n8FTI.png

DVPrtWy.png

vSmfw1y.png

The rest of the hardware is nothing fancy, a Soyo 7VCA motherboard (which has a whole lot of "quirks and features" by itself, mostly quirks tho) and ram is a mix of value Kingston pc100 and 6 ns Twinmos.

I could guess that maybe Sandra favours raw clock speed over anything else, but if that were the case the P4 would be miles ahead.

Sorry for the cheesy title but it was the least worst thing that came to mind.
Andrea

Reply 2 of 10, by flupke11

User metadata
Rank Member
Rank
Member

Try a pi calculation test and see how that fares. The Coppermine is a good architecture. The P4 only really shows its teeth at SSE(2) instructions and at high FSB speeds.

Your Cellie is probably a late one, and capable of much higher stock speeds than the one it got on its final label. A PIII at the same FSB/core speed should logically be even faster. And the Tualatin, well, that's just the best thing ever to have come out of Intel's factories in the early years of the Millennium.

Reply 4 of 10, by maxtherabbit

User metadata
Rank Oldbie
Rank
Oldbie

yeah the cache deficiency is not going to manifest in dhrystones, whetstones or memory mandbidth tests. You have a strong CPU no doubt but it's no surprise that the extra ~150MHz puts it ahead of the PIII 1G

Reply 5 of 10, by andrea

User metadata
Rank Member
Rank
Member

Thanks for the suggestions, i'll try with a 1G P3 (I know I have it somewhere, just have to find it) with a fastish VGA so that the bottleneck will be the CPU and come back later with results.

Have a nice day,
Andrea

Reply 6 of 10, by alvaro84

User metadata
Rank Member
Rank
Member

You made me curious. You've eliminated one of the Celeron's bottlenecks here, the lower FSB. The other, the halved cache, of course, remains. Let's see how it fares against the full fledged P3.

Shame on us, doomed from the start
May God have mercy on our dirty little hearts

Reply 7 of 10, by dionb

User metadata
Rank l33t
Rank
l33t

That deficit you mention came from the very high multiplier vs low cache on the later models. This one is a Celeron with relatively low multiplier, so the lack of cache doesn't bite so hard. You have a really, really fast Coppermine CPU with good FSB - but low cache. A P3-1133 would beat it of course, but I'm not surprised if it surpasses a P3-1000EB and it definitely would beat a P3-1000E.

Still, Q3A would be a better heavy test.

Last edited by dionb on 2020-05-14, 15:00. Edited 1 time in total.

Reply 9 of 10, by andrea

User metadata
Rank Member
Rank
Member

I've been doing some tests, but the motherboard decided to play Abit and stop posting.
Anyway you people were right, while the Celeron may be faster in Sandra, 3D stuff is a different story.

I'll leave the board without power and jumpered to clear cmos for a few hours and then try to see if it behaves.
I am missing just a 3dmark run and then I'll post the results

Reply 10 of 10, by andrea

User metadata
Rank Member
Rank
Member

Ok, I'm back. I got the motherboard working after disassembling the socket and pouring some alcohol over the contacts inside.

All the following benchmarks have been done using a Radeon 8500@9100 275/275 and 512 MB of RAM running 1:1 with the FSB at 2-2-2-5 timings. OS was a clean and slighty tweaked XP SP3 install.
A true comparison would have required a P3 1133, which I don't have. Instead, I've used a P3 1000 EB two times, at first set at the same bus speed as the Celeron, and then at the same clock speed* (*almost the same, due to choices the PLL chip offered me).

System setup:
Chipset VIA 694X/686A with 4in1 4.51v
Radeon 9100 with Catalyst 5.8
Yamaha 724 with drivers 5244
Intel PRO 100+ with drivers 8.0.47

1) Celeron at 1147 MHz (8.5 * 135)

SKOheIt.png

Sandra 2002:

W8mTryG.png

SLgh95J.png

fla01i7.png

Super PI 1M:

URQPnVK.png

3DMark 2001 SE (default settings):

UOwn9dX.png

Quake 3 timedemo (1024x768 32 bit colour): 94.4 fps

================================================================================

2) Pentium 3 at 1012 MHz (7.5 * 135) - same bus speed

GHnhDiu.png

Sandra 2002:

oWOQBJO.png

d4yoWej.png

hfKT5zm.png

Super PI 1M:

6fHcDpD.png

3DMark 2001 SE (default settings):

TQOHrS2.png

Quake 3 timedemo (1024x768 32 bit colour): 103.3 fps

================================================================================

3) Pentium 3 at 1125 MHz (7.5 * 150) - same clock speed

OBf8TZT.png

Sandra 2002:

EYtPVr2.png

9ENOZFS.png

EXEKLVn.png

Super PI 1M:

https://i.imgur.com/9t9Sg43.png

3DMark 2001 SE (default settings):

Low9va8.png

Quake 3 timedemo (1024x768 32 bit colour): 113.3 fps

================================================================================

I've come to the conclusion that it's the P3 that's faster than I thought it would be and, by extension, so is the Celeron. But no matter what the lesser cache does have quite an impact.
And I guess Sandra results have to be taken cum grano salis.

Andrea