To switch, or not to switch

[xenos500]

hey everyone,

I have had my P4 2.53 for a while now, and I got a new MB and ram for it (IC7 + 512DDR) a month ago.

Over the weekend I got a cheapo Celeron 2.0 and dropped it into my old MB and ram (P4t533 + RamBus1066) to make a 'secondary' computer.

As a given, i bumped the celeron up to 133/533 FSB....turning it into a 100%stable 2.66ghz
Then I upped the voltage .03v and now it runs at 140/560fsb ...turning it into a 100% stable 2.8ghz

Basicly...my secondary computer is kicking my primary systems a$$ off in every benchmark I can find.

should I put the celeron in my #1 system with the slightly better ram...better MB and MUCH better video card?

this seems too good to be true....the CPU was only $50?I know the cache is 1/4th the size....but that doesnt seem to hurt it in benchmarks.

What should i do?
my goal isnt to balance resources...I want my primary system to be as fast as possible...and use the left overs in a secondary sys.

thanks alot

 

[Pentium]

NOOOOO!!!

the 400k of cache difference will hinder performance!!!


Overclock the P4 2.53!!!

 

[Duvie]

What benches are you doing??? I mean in most celeron test they had to oc them to 3ghz and the things would still lose to 2.2-2.4ghz p4's. Celerons are generally junk unless you just web surf and word process. If this thing is beating your first system in multimedia, rendering, encoding and gaming apps then you have somehow f'd up the first system cause it should not happen.

Granted the rdram is a nice system the fact is the dual channel ddr (if you are actually run it) should outperform the rdram, and the celeron is lacking l1 and l2 cache compared to its brother the p4b and in the p4 architecture it is shown te p4 takes advaantage of the the added 12 cache. Heck it jumped near 10-15% when gong from 256kb willamette core to the northwood 512kb one.

Let me guess!!! Got 1 stick of 512mb ddr in that abit ic7 mobo of 2???


The celeron should never beat the p4b period...Unless you have neutered the DCDDR mobo, using like a 5400rpm or less hDD versus faster 200rpm in the other...As well as just plain use better vid component in secong system and testing it in gaming....

Just shouldn't happen!!!

 

[madcow235]

Unless you are using pure clock based benchmarks the Celeron should never win.

 

[lordtyranus2]

Have you overclocked the p4? It looks like you are comparing overclocked Celeron to stock p4, which isn't fair.

 

[jiffylube1024]

Originally posted by: lordtyranus2
Have you overclocked the p4? It looks like you are comparing overclocked Celeron to stock p4, which isn't fair.

No, even still a 2.53 GHz P4 will crush the sh*t out of a 2.8GHz Celeron. Something is seriously amiss, and these "benchmarks" must be screwed up.

Try running SiSoft Sandra's CPU benchmark. The P4 should destroy the Celeron. Try some newer games (UT2k3 would be a good choice). Benchmark with both CPU's and you will see the P4 blows the Celeron out of the water.

 

[DivideBYZero]

The Celeron is an Asthmatic Ant with three bags of very heavy shopping. The P4 is a lean mean long distance runner.

Don't get fooled.

 

[Soulkeeper]

What is the P4 overclocking to ?
it should overclock about as well as the celeron

 

[xenos500]

Originally posted by: Duvie

Let me guess!!! Got 1 stick of 512mb ddr in that abit ic7 mobo of 2???

no cheese head, I have 2 256's

but I put them both in the first controller because I dont like dual channel.

 

[xenos500]

Originally posted by: lordtyranus2
Have you overclocked the p4? It looks like you are comparing overclocked Celeron to stock p4, which isn't fair.

my p4 doesnt over clock well....its running at 2.66 now though....seems ok

 

[xenos500]

let me get some actuall benchmark results here....check later

 

[Duvie]

Originally posted by: xenos500

Originally posted by: Duvie

Let me guess!!! Got 1 stick of 512mb ddr in that abit ic7 mobo of 2???

no cheese head, I have 2 256's

but I put them both in the first controller because I dont like dual channel.

That is what I figured cheese-dick!!! You have confirmed my suspicion you are a moron!!! I got the sense you didn't know what you were doing real quick....

"I don't like dual-channel" LOL.....I guess you don't like running it the way it was intended and the way to maximize the processors ability. If you were ocing and somehow it was limiting your max OC I may be able to understand this, but this is out in noobie left field!!!

Go ahead and sell the p4b and kee the celeron...You deserve lackluster performance....


I will be laughing about this one for awhile!!!! HAHAHAHAHAHAHA

 

[xenos500]

uh, yea cheese <insert body part here>, I was joking.

heres the results of my recent comparison testing:

P4 2.53 @ 2.66
SiSoft Sandra 2004
CPU Arithmatic
Dhrystone ALU 6824 MIPS
Whetstone FPU 1966 MFLOPS + iSSE2 3593 MFLOPS
CPU MultiMedia
Integer iSSE2 16510 it/s
Float iSSE2 19529 it/s
MemoryBench
Int Buffered iSSE2 3561 MB/s
Float Buffered iSSE2 3561 MB/s

Adobe Photoshop 6.0
Apply Cutout filter to 7052_a2000f118.jpg (434kB 2824x2092 1296dpi)
15.10 seconds
Apply StainedGlass filter to 7052_a2000f118.jpg (434kB 2824x2092 1296dpi)
18.60 seconds

WinRAR
Compress 90 .jpg images totaling 77.5MB into .RAR with normal compression
49 seconds






P4 Celeron 2.0 @ 2.86
SiSoft Sandra 2004
CPU Arithmatic
Dhrystone ALU 7165 MIPS
Whetstone FPU 2110 MFLOPS + iSSE2 3853 MFLOPS
CPU MultiMedia
Integer iSSE2 17703 it/s
Float iSSE2 20943 it/s
MemoryBench
Int Buffered iSSE2 3091 MB/s
Float Buffered iSSE2 3433 MB/s

Adobe Photoshop 6.0
Apply Cutout filter to 7052_a2000f118.jpg (434kB 2824x2092 1296dpi)
21.2 seconds
Apply StainedGlass filter to 7052_a2000f118.jpg (434kB 2824x2092 1296dpi)
19.9 seconds

WinRAR
Compress 90 .jpg images totaling 77.5MB into .RAR with normal compression
1minute 16 seconds


Summary

Celeron scores better in synthetic benchmarks
P4 tore faster clocked celeron a new one in Photoshop and winRAR


For complete system specs, see my system rigs
Sprygg = P4
The-Bus = Celeron

For screen shots
go here:
http://sprygg.ath.cx:8080

 

[Boonesmi]

Originally posted by: Duvie

Originally posted by: xenos500

Originally posted by: Duvie

Let me guess!!! Got 1 stick of 512mb ddr in that abit ic7 mobo of 2???

no cheese head, I have 2 256's

but I put them both in the first controller because I dont like dual channel.

That is what I figured cheese-dick!!! You have confirmed my suspicion you are a moron!!! I got the sense you didn't know what you were doing real quick....

"I don't like dual-channel" LOL.....I guess you don't like running it the way it was intended and the way to maximize the processors ability. If you were ocing and somehow it was limiting your max OC I may be able to understand this, but this is out in noobie left field!!!

Go ahead and sell the p4b and kee the celeron...You deserve lackluster performance....


I will be laughing about this one for awhile!!!! HAHAHAHAHAHAHA

hehe i think he was just being sarcastic when he said "but I put them both in the first controller because I dont like dual channel"

 

[mrgoblin]

WHy you didint oc ure 2.53 in the first place begs the question why did you buy the celeron but aside from that, think about what your saying?

 

[Shimmishim]

your tests go to show that synthetic benches are crap...

thanks for actually doin' some REAL tests

it's refreshing to see people do real aps rather than that sisoft junk...

 

[xenos500]

Originally posted by: Shimmishim
your tests go to show that synthetic benches are crap...

thanks for actually doin' some REAL tests

it's refreshing to see people do real aps rather than that sisoft junk...

Yea, the synthetic benchmarks had me beliving that my celeron was outpreforming my P4 at nearly one quarter the cost.

It quickly became clear that the synthetic benchmarks were crap. I wish someone could standardize some actual benchmark tests.....similar to what I did. Photoshop is an excelent tool that many people actually have, and PS offers the ability to time tasks with tenth of a second accuracy. I think winZIP and winRar are also good....as well as video encoding/decoding.

Im going to start a new thread in general hardware with some hosted files for people to run benchmarks and post results from actual applications. It would be great to have some benchmarks that didnt come from 3dmark or some other pretend benchmark.

 

[Cerb]

"The Celeron is an Asthmatic Ant with three bags of very heavy shopping."
...that's going to me sig

 

[xenos500]

Originally posted by: Cerb
"The Celeron is an Asthmatic Ant with three bags of very heavy shopping."
...that's going to me sig

ants can carry many times their body weight.