Real-world impressions of going from P4 to AMD64 3200+...**UPDATE** Up and Running...FINALLY :)

[Markfw]

Originally posted by: Markfw900
NFS4, what about your impressions ? Does it seem noticably faster than your P4 3.0 ? That is what I am most interested in (the original thread title)

Just a bump.

 

[flexy]

i am getting kinda suspicious not reading too much about benchies etc. here...sandra scores...good...well....i mean it's AMDs new processor...so..we wanna know how fast it is
Btw. i still think a comparison to Barton XP3200+ would be interesting...would very clearly show the improvement (IF ANY) to XP/Barton and the A64 running at the same speed...

 

[cmdrdredd]

Originally posted by: flexy
i am getting kinda suspicious not reading too much about benchies etc. here...sandra scores...good...well....i mean it's AMDs new processor...so..we wanna know how fast it is
Btw. i still think a comparison to Barton XP3200+ would be interesting...would very clearly show the improvement (IF ANY) to XP/Barton and the A64 running at the same speed...

well, from the sandra scores it looks like I won't be recommending a Athlon64

 

[Cerb]

Originally posted by: cmdrdredd

Originally posted by: flexy
i am getting kinda suspicious not reading too much about benchies etc. here...sandra scores...good...well....i mean it's AMDs new processor...so..we wanna know how fast it is
Btw. i still think a comparison to Barton XP3200+ would be interesting...would very clearly show the improvement (IF ANY) to XP/Barton and the A64 running at the same speed...

well, from the sandra scores it looks like I won't be recommending a Athlon64

Real benches matter. Sandra scores are generally BS.

 

[cmdrdredd]

Originally posted by: Cerb

Originally posted by: cmdrdredd

Originally posted by: flexy
i am getting kinda suspicious not reading too much about benchies etc. here...sandra scores...good...well....i mean it's AMDs new processor...so..we wanna know how fast it is
Btw. i still think a comparison to Barton XP3200+ would be interesting...would very clearly show the improvement (IF ANY) to XP/Barton and the A64 running at the same speed...

well, from the sandra scores it looks like I won't be recommending a Athlon64

Real benches matter. Sandra scores are generally BS.

not really...sandra is quite accurate for measuring overall speed if the settings are the same across the board. In such a case the fastest CPU is the fastest CPU and no video driver, game version, services running etc will change your score

sounds to me like you're just covering up the fact that in pure power the Athlon64 may not be as fast as once thought.

 

[Cerb]

Originally posted by: cmdrdredd

Originally posted by: Cerb

Originally posted by: cmdrdredd

Originally posted by: flexy
i am getting kinda suspicious not reading too much about benchies etc. here...sandra scores...good...well....i mean it's AMDs new processor...so..we wanna know how fast it is
Btw. i still think a comparison to Barton XP3200+ would be interesting...would very clearly show the improvement (IF ANY) to XP/Barton and the A64 running at the same speed...

well, from the sandra scores it looks like I won't be recommending a Athlon64

Real benches matter. Sandra scores are generally BS.

not really...sandra is quite accurate for measuring overall speed if the settings are the same across the board. In such a case the fastest CPU is the fastest CPU and no video driver, game version, services running etc will change your score

sounds to me like you're just covering up the fact that in pure power the Athlon64 may not be as fast as once thought.

1. Services running should be the same for any set of tests on any review.
2. Pure power at what? Sandra's tests are too isolated. Business Winstone gives you the overall performance at things people do. Other benchmarks do the same. All Sandra can really do is show whether you've got something set wrong. If your HD isn't within 10% or so of their reference, you might want to look at your drivers and such. If your int, float, sse, whatever are less, you might have it underclocked, whoops. Otherwise it really doesn't say that much, except for raw power at tasks that are for more isolated than anything people actually do. If the A64's Sandra scores being a bit low mean it's lesser overall, why is it competitive and in many cases better than the P4 in the gaming benches? Shouldn't they be representative of what you're getting in Sandra?

 

[cmdrdredd]

perhaps we need memory behchmarks? It may be that the CPU is not where the real performance comes from...

remember how the P4 dominated the Athlon in gaming because of memory bandwidth?

 

[Cerb]

Originally posted by: cmdrdredd
perhaps we need memory behchmarks? It may be that the CPU is not where the real performance comes from...

remember how the P4 dominated the Athlon in gaming because of memory bandwidth?

Yup, and it's no suprise that it's the Opteron that can compete with it in media encoding due to that.

 

[cmdrdredd]

Originally posted by: Cerb

Originally posted by: cmdrdredd
perhaps we need memory behchmarks? It may be that the CPU is not where the real performance comes from...

remember how the P4 dominated the Athlon in gaming because of memory bandwidth?

Yup, and it's no suprise that it's the Opteron that can compete with it in media encoding due to that.

I'd really like to compare some memory benchies.

 

[Viper96720]

Maybe VIA is the reason for the scores. The Anandtech review used a nforce 3 and A64 did well in pretty much everything

 

[Duvie]

You know it has been almost 4 days since NFS4 has updated this and given us some benchies...Do we need a search party???

I thought the memory controller on die was to take away the real difference in mobo manufacturers. It should negate VIA's historic crappy memory performance. The pci and agp bus may be a different story though...

 

[Budman]

Originally posted by: Duvie
You know it has been almost 4 days since NFS4 has updated this and given us some benchies...Do we need a search party???

I thought the memory controller on die was to take away the real difference in mobo manufacturers. It should negate VIA's historic crappy memory performance. The pci and agp bus may be a different story though...

I am thinking Brandon's having buyer's remorse,He had a stable & fast P4 setup & now he's Beta testing. :disgust:

 

[flexy]

Originally posted by: Budman

Originally posted by: Duvie
You know it has been almost 4 days since NFS4 has updated this and given us some benchies...Do we need a search party???

I thought the memory controller on die was to take away the real difference in mobo manufacturers. It should negate VIA's historic crappy memory performance. The pci and agp bus may be a different story though...

I am thinking Brandon's having buyer's remorse,He had a stable & fast P4 setup & now he's Beta testing. :disgust:

lol yeah i can see that.... on the other hand i'd like to see that SETI benches....hmm...

 

[NFS4]

Originally posted by: Budman

Originally posted by: Duvie
You know it has been almost 4 days since NFS4 has updated this and given us some benchies...Do we need a search party???

I thought the memory controller on die was to take away the real difference in mobo manufacturers. It should negate VIA's historic crappy memory performance. The pci and agp bus may be a different story though...

I am thinking Brandon's having buyer's remorse,He had a stable & fast P4 setup & now he's Beta testing. :disgust:

Isn't it funny how people try to ANSWER for you???

I've been away from my computer at home since Wednesday morning. I left Raleigh to go to Charlotte Wednesday to go see my GF and spent Wed and Thurs with her. Then I came back to Burlington last night to spend time with my parents since I don't see them much. And then my GF is coming up here today to spend the day with me. And I won't be back to my computer in Raleigh until Sunday. So right now, I'm typing on my mom's 900MHz IBM T22 laptop.

I'm sorry, but I have things that are important to me than my computer ya know, so I'm sorry if I'm not spitting out benchmarks 24-7. I understand that you guys want to hear performance figures, but there are other things on my mind that are of greater priority.

Anyway, as for Budman's OTHER comments...the platform has been rock solid since I got everything installed. It turns out that ALL of the problems I was having were attributed to the Windows XP CD as all of my problems were related to booting (someon explained it about 10-15 posts up). Probably a BIOS update will fix that. And even so, that's not a VIA issue...that an FIC BIOS issue. Seems as though everyone is always quick to bash VIA around here, but everyone seems to forget NVIDIA's issues with nForce IDE performance or buggy APU drivers...

Regardless, my system has been scorching fast and I love it (from the time I DID get to spend with it since I built it). I'm now shopping around for a Radeon 9800NP to throw in there since I just got back $200 in rebate checks that I had forgotten about That's how I pay for all of my hardware anyway. Hot Deals + eBay + Rebate Checks

 

[dguy6789]

Originally posted by: ahsia
I always say unless you are running applications that require insane amount of processing power, there is really no need for the TOP of the line processor. I can imagine something like a 2.4C P4 or a XP 2600+ can handle anything you throw at it. In the end, it is how much memory, and how fast your hard drive that will make any real world differences.

not even that, heck, an athlon xp 1600 with 512MB of ram and a nforce 2 and radeon 9700 pro will handle anything out today just fine at highest settings.

 

[dguy6789]

Originally posted by: ahsia
I always say unless you are running applications that require insane amount of processing power, there is really no need for the TOP of the line processor. I can imagine something like a 2.4C P4 or a XP 2600+ can handle anything you throw at it. In the end, it is how much memory, and how fast your hard drive that will make any real world differences.

not even that, heck, an athlon xp 1600 with 512MB of ram and a nforce 2 and radeon 9700 pro will handle anything out today just fine at highest settings.

 

[Budman]

Sorry Brandon I didnt know you were busy.


Yes friends & family is way way way more importand that any peice of hardware,I hope you have fun & enjoy your time away.

 

[Megatomic]

Originally posted by: NFS4
Seems as though everyone is always quick to bash VIA around here, but everyone seems to forget NVIDIA's issues with nForce IDE performance or buggy APU drivers...

Not I, I live with NVidia's crappy IDE and APU drivers and dream of new drivers being released that magically fix it all for me. I tell ya, if the VIA KT880 does well I may ditch my Abit NF7-S R2 for an Abit KV7II. :|

 

[ssanches]

Here's a nice mATX AMD64 mobo from Gigabyte. Maybe I shall take the AMD64 plunnge sometime soon

 

[pelikan]

Congratulations NFS4. Nice new system.

 

[NFS4]

Originally posted by: pelikan
Congratulations NFS4. Nice new system.

Thanks. I'll be sure to give you guys overclocking results when I get back

 

[fkloster]

I am waiting for N4S4's return. I have, basically, resurrected his old P4 rig (3.2 ghz now...) and am ready to test it against his new Athlon64 3200+ rig just for kicks... Brandon..this is what I have so far:

6,650 3dMarks build 330

50,453 AquaMark3

Sandra Standard v. 2002.6.8.97

CPU Arithmetic Benchmark:

Dhrystone ALU = 7,421 Mips
Whetstone FPU = 6,496 Mflops

CPU Multimedia Benchmark:

Integer = 14,921 it/s
Floating Point = 22,662 it/s

Filesystem Benchmark:

Drive Index = 67,790 mb/s

Memory Bandwidth Benchmark:

Ram Int Buffered = 6,145 mb/s
Ram Float Buffered = 6,173 mb/s

Cache & Memory Benchmark:

2kb-----34,204 mb/s
4kb-----34,978 mb/s
8kb-----28,355 mb/s
16kb-----24,797 mb/s
32kb-----25,366 mb/s
64kb-----25,381 mb/s
128kb-----25,808 mb/s
256kb-----23,542 mb/s
512kb-----19,474 mb/s
1mb-----2,985 mb/s


I know this is a bunch of synthetic crap... but lets get our heads together and isolate some kind of some 'real world' tests too ok?

 

[NFS4]

Originally posted by: fkloster
I am waiting for N4S4's return. I have, basically, resurrected his old P4 rig (3.2 ghz now...) and am ready to test it against his new Athlon64 3200+ rig just for kicks... Brandon..this is what I have so far:

6,650 3dMarks build 330 (N4S4's = 3,617)
cpu score = 740 (N4S4's cpu score = 720)

50,453 AquaMark3

Sandra Standard v. 2002.6.8.97

CPU Arithmetic Benchmark:

Dhrystone ALU = 6,021 Mips (N4S4's ALU = 8,308 Mips)
Whetstone FPU = 6,292 Mflops (N4S4's FPU = 4,085 Mips)

CPU Multimedia Benchmark:

Integer = 14,897 it/s (N4S4's Integer = 11,213 it/s)
Floating Point = 22,662 it/s (N4S4's Floating Point = 12,319 it/s)

Filesystem Benchmark :

Drive Index = 67,790 mb/s

Memory Bandwidth Benchmark:

(module would not run?)

Cache & Memory Benchmark:

2kb-----34,204 mb/s
4kb-----34,978 mb/s
8kb-----28,355 mb/s
16kb-----24,797 mb/s
32kb-----25,366 mb/s
64kb-----25,381 mb/s
128kb-----25,808 mb/s
256kb-----23,542 mb/s
512kb-----19,474 mb/s
1mb-----2,985 mb/s

I know this is a bunch of synthetic crap... but lets get our heads together and isolate some kind of some 'real world' tests too ok?

(1) Fkloster, you're mixing my benches from 2003 and 2004 Sandra together
(2) Remember, you're using a Radeon 9800 Pro....I'm using a Radeon 9600 Pro (in 3d app benchmarks)

I've got to run some HD benches, be back in about 10 mins

 

[flexy]

so..in other words his old rig was faster ? uhmm...yes...cough...ok. Let's stop this overclocked CPU benchies....they are pointless On the other hand...look, you had to run the P4 at 3.2Ghz (!) to get about the same performance as A64 at 2Ghz....gulp...what does that tell us ?

Btw. i dont *think* that the A64 has a high OC potential..but who knows.

Isn't there ANY special area where the A64 has a significant performance lead ? Logic says there must be one.....
(Btw2: I think it would be time to get *somewhat* concerned in case there is no big OC capability with the A64 - and on the other hand it would be easy to OC the P4 so it beats A64 in CPU benchies etc. ) Then i really woudlnt see a point in A64. Different, however, if there is some space for some nice A64 OC - then the whole thing gets interesting...

 

[flexy]

(1) Fkloster, you're mixing my benches from 2003 and 2004 Sandra together
(2) Remember, you're using a Radeon 9800 Pro....I'm using a Radeon 9600 Pro (in 3d app benchmarks)

I've got to run some HD benches, be back in about 10 mins

oups..fkloster...tsk tsk tsk...

3dmark is of course worthless right now - so get a 9800 whatever and give as the score then...lol

HD benches...well..nice...but they dont really say anything about the A64 ? I think Seti would be good for comparison.
Sandra should be ok if its the same versions