My new Intel rig => how does it compare to AMD 64 rigs ?

[BEIF]

Hello

Just put together a new Intel rig as follows:-

2.6C @ 3.315 @ 1.60v in BIOS gievs me 1.664v at idle in Win & 1.584 under load in Win
Zalman AlCu - A 7000
13 x 255 mhz FSB
1:1 memory @ 3/4/4/8 @ 2.75v
Hyperthreading **DISABLED**
2 x 512 Hynix D43 DDR
Asus P4P 800 BIOS 1016
All Prime stable


[ CPU will go higher except for that damn P4P800 BIOS bug whereby the 1.625 and 1.650 cpu voltage settings dont allow the system to boot up - 1.675v works fine, but takes the "real" cpu voltage to 1.74v - too high for my liking - *really* hope they fix this bug in a new bios]

Anyhow - the following are my basic benchmark results - *all* benchmarks done with Hyperthreading turned OFF in the BIOS:-

Super Pi
1M 42s
2M 1m 34s
4M 3m 29s
8M 7m 32s

CTIAW 2.7
6514mb ? partial PAT enabled

Sandra SP1 - CPU Arithmentic
DS 8573
WS 2422 / 4466 MFLOPS

Sandra SP1 - Multimedia Arithmentic
Integer 20564
FP 25691

Sandra SP1 - Memory Bandwidth
RAM Int 5845
RAM Float 5851

How do the above figures stack up against say an AMD 64 3200+ or 3400+ ?

TIA,
Ben

 

[Markfw]

There are plenty of benchmarks all over the web. My Athlon64 3000+ at 2.1 probably equals or bests (in most cases) your 3.3 ghz in everything most people care about. If not, read the benchmarks..... And when 64-bit hits mainstream in 6 months, astalavista baby.......

If you like your system fine, but don't post flamebait and expect any reasonable answers...

 

[AWhackWhiteBoy]

my barton at 240x10(2.4ghz) gets the same superpi score.

 

[Shimmishim]

Beif,

It's not a bios bug that the voltage goes up that high at 1.65 or 1.675 volts...

when companies make boards, the voltage has to be within a certain percentage of the set voltage...

and it is within the spec...

have you ever monitored the voltage under load?

it drops like crazy!!!

there is something called a "vdroop" mod that prevents this from happening.

if you want more info check out xtremesystems.org or PM me.

 

[BEIF]

Thanks AWhackWhiteBoy - much appreciated

Markfw900 ... thank you too for your 3000+ @ 2.1 info - also much appreciated

I was/am simply seeking some honest advice / opinion and this is one of the few forums where people can and do provide that becuase it strikes me that the people here actually "do" know what they are talking about [ unlike many other forums ] - if my post came across as somehow "fishing" for *anything* other than genuine advice or comment, then I apologize .... I will chosse my words more carefully next time

Ben

 

[BEIF]

Thanks Shimmishim

The bug / problem with my [ and many other P4P800 owners ] board is that if I set the bios CPU voltage setting to either 1.6250v or 1.650v, the system will simply not boot up in any shape or form.

If I select 1.675v, it will work fine, but my idle "real" world voltage as measured in Windows by Asusprobe will be around 1.74v <-> 1.76v at idle - that figure, on air, scares the you-know-what out of me.

Reading between your lines, are you suggesting that the load / useage voltage reading is more important than the idle voltage reading ?

Ben

 

[Markfw]

OK, no more antagonism. But synthetic benches are crap. Just ask for something I can run, and I will try. But really, flames aside, Intel is hurting right now, and can't compete in anything most people care about. I can set the computer to encode overnight, and the 5 minutes it will save on a 2 hour home movie don't impress me.

Give be something real to run and I will.....

 

[3LEMENT0]

don't have sandra but weird thing is our super PI score is the same....mine's just a sec faster at 4M and 8M. AXP-Mobile @2532Mhz. Enjoy your rig....

 

[oldfart]

Curious...why is Hyperthreading disabled?
That Vcore is pretty high as it is. I wouldn't go any higher anyway.
3-4-4-8 timings is pretty slow. Can that be improved?

And yes, skip the synthetic bench stuff. Bench some applications that people actually use.

 

[imported_michaelpatrick33]

Enjoy your rig man it is definitely fast enough! I still gaze at my Pentium II 450 that I paid 2700 dollars years ago with fondness!

 

[oldfart]

Originally posted by: Markfw900
OK, no more antagonism. But synthetic benches are crap. Just ask for something I can run, and I will try. But really, flames aside, Intel is hurting right now, and can't compete in anything most people care about. I can set the computer to encode overnight, and the 5 minutes it will save on a 2 hour home movie don't impress me.

Give be something real to run and I will.....

5 mins on a 2 hr home movie? Its more than that. What software are you using? The two I mainly use are Pinnacle Studio 9 and TMPGEnc.
THG benches (I know, I dont like Toms either, but he does use these apps for benches).

P4 3.2 is ~ 20% faster the A64 3200+ in Studio 9 and TMPGEnc. That is WAY more than 5 min on a 2 hr movie.

For games in that same article
Wolf ET, both ~ the same (148 Vs 150 FPS)
Serious Sam again, 101 Vs 106 FPS
Comanche 4 66 Vs 68 no difference
UT2K3, A64 wins here 278 Vs 249 about a 13% advantage
Splinter Cell A64 also wins 85 Vs 71
X2 A64 157 Vs 148
Aquamark same performance

For gaming they both have very playable frame rates in every bench. Many were ~ a tie. In none did the A64 have a 20% advantage like the P4 has in encoding. It depends on what is important to you.

For me, 101 Vs 106 or 278 Vs 249 FPS doesn't matter. You cant tell the difference. 6 Vs 7 Hr encode time is not all that big a deal either.

I think more is made of one Vs the others advantage or disadvantage than is reality. Either one is very fast no matter how you slice it.

 

[BEIF]

Thanks again Markfw900

Thanks 3LEMENT0

oldfart

=> I have to run H/T disabled as one of my hioh end Digital Audio applications is not "thread-safe" - the effect being that with H/T turned on, it causes quite wild CPU useage spiking which leads to multi-track audio playback problems - it is a known problem with that softweare, and hpefully will be
"addressed" soon.

=> I'm not likely to try and go past 1.664v real / actual on the vCore either

=> my DDR Ram is generic with the "classic" Hynix D43 PC3200 chips - it will happily work at up to 269mhz at 1:1 with 2.75v - I have tested it in another P4 machine at that speed and it runs fine all day long - memtest, prime etc .... problem is that (i) my current 2.6C tops-out at 255mhz and (ii) the ram will only run at sub 3/4/4/8 settings at no higher than 210mhz - as such, I'd much rather have the 255fsb at 1:1 than 210 at tight timings

=> yep - 3.3 gig is pleanty fast

michaelpatrick33

Thanks too - I will enjoy - $2700 for a PII 450 - I know *EXACTLY* what you mean !!!


Thanks all for your comments.

Ben

 

[oldfart]

Your setup is similar to mine. I also run P4 2.6 @ 255 FSB. I use an Abit IC7 board. Abit undervolts Vs overvolt of Asus. I'm @ 1.575 BIOS setting, 1.55 actual. My ram is more suited to low latency. I run 5:4 ratio with 2-2-2-6 timings. The performance diff between 1:1 loose and 5:4 tight timings is pretty small with the 5:4 tight winning by a small margin. I dont think that your ram can do the 2-2-2 timings anyway, just like mine could never get to DDR510 @ any timings.

 

[Markfw]

Originally posted by: oldfart

Originally posted by: Markfw900
OK, no more antagonism. But synthetic benches are crap. Just ask for something I can run, and I will try. But really, flames aside, Intel is hurting right now, and can't compete in anything most people care about. I can set the computer to encode overnight, and the 5 minutes it will save on a 2 hour home movie don't impress me.

Give be something real to run and I will.....

5 mins on a 2 hr home movie? Its more than that. What software are you using? The two I mainly use are Pinnacle Studio 9 and TMPGEnc.
THG benches (I know, I dont like Toms either, but he does use these apps for benches).

P4 3.2 is ~ 20% faster the A64 3200+ in Studio 9 and TMPGEnc. That is WAY more than 5 min on a 2 hr movie.

For games in that same article
Wolf ET, both ~ the same (148 Vs 150 FPS)
Serious Sam again, 101 Vs 106 FPS
Comanche 4 66 Vs 68 no difference
UT2K3, A64 wins here 278 Vs 249 about a 13% advantage
Splinter Cell A64 also wins 85 Vs 71
X2 A64 157 Vs 148
Aquamark same performance

For gaming they both have very playable frame rates in every bench. Many were ~ a tie. In none did the A64 have a 20% advantage like the P4 has in encoding. It depends on what is important to you.

For me, 101 Vs 106 or 278 Vs 249 FPS doesn't matter. You cant tell the difference. 6 Vs 7 Hr encode time is not all that big a deal either.

I think more is made of one Vs the others advantage or disadvantage than is reality. Either one is very fast no matter how you slice it.

Since I don't like to argue small points, I actually have to agree. There is not that much difference, except the Athlon64 wins more in games (real-time, counts more by me) and it runs cooler and cost less (last I looked). So, all in all, I get the Athlon64.

 

[cubanx]

Hi BEIF.

Here are the Sandra results from my 3 day old 3200+ A64 for comparison.

Sandra SP1 - CPU Arithmentic
DS 10974
WS 4005/ 5132 MFLOPS

System Specs:
A64 3200+ @ 2.53GHZ (230fsbx11)
MSI K8N Plat. NForce 3 250GB
1 GB Ram
ATI 9800Pro

 

[Shimmishim]

Originally posted by: BEIF
Thanks Shimmishim

The bug / problem with my [ and many other P4P800 owners ] board is that if I set the bios CPU voltage setting to either 1.6250v or 1.650v, the system will simply not boot up in any shape or form.

If I select 1.675v, it will work fine, but my idle "real" world voltage as measured in Windows by Asusprobe will be around 1.74v <-> 1.76v at idle - that figure, on air, scares the you-know-what out of me.

Reading between your lines, are you suggesting that the load / useage voltage reading is more important than the idle voltage reading ?

Ben

i think both are important... but load is probably more important especially for those overclocking.

i've owned a p4p800 rev 2 before... and i understand what you mean by load and unload voltages...

this isnt' a simple bios fix... it's an actual hardmod fix...

your idle voltage will still be a bit high... but your load voltage will be pretty constant compared to w/o the mod.

 

[charloscarlies]

Since I don't like to argue small points, I actually have to agree. There is not that much difference, except the Athlon64 wins more in games (real-time, counts more by me) and it runs cooler and cost less (last I looked). So, all in all, I get the Athlon64.

Since when did an A64 cost less than a 2.6C? The P4 is $50-80 cheaper.

 

[BEIF]

Thanks oldfart.

Shimmishim re: the P4P800 BIOS voltage - what I dont understand is that all the settings *before* 1.600v in the bios [ ie: 1.575, 1.550, 1.525 ] work fine * &amp; * all the settings *after* 1.650v in the bios [ ie 1.675v, 1.700v and up ] all work fine.

As such then, I cant see any *hardware* reason why the 1.625v and 1.650v settings stop the machine from booting up !?!? To my mind, it has to be some sort of bad or corrupted very-low level bios code ?

Cubanx - thanks for the figures and specs - man, now that is a nice machine - if I am not mistaken, you effectively have a AMD 64 3700+ machine there, except yours is quicker because a "real" 1mb cache 3700+ cpu - if there ever was to be such a thing - would likely run at 12.5 x 200 = 2500mhz, whereas you are at a higher fsb and lower multiplier - [ 11 x 230 ] which gives a higher processing "throughput" - well done !!

Cubanx - what sort of cooling are you using ? - and what are your temps at idle and full load in a closed case ?

Thanks all,
Ben

 

[cubanx]

Originally posted by: BEIF
Cubanx - thanks for the figures and specs - man, now that is a nice machine - if I am not mistaken, you effectively have a AMD 64 3700+ machine there, except yours is quicker because a "real" 1mb cache 3700+ cpu - if there ever was to be such a thing - would likely run at 12.5 x 200 = 2500mhz, whereas you are at a higher fsb and lower multiplier - [ 11 x 230 ] which gives a higher processing "throughput" - well done !!

Cubanx - what sort of cooling are you using ? - and what are your temps at idle and full load in a closed case ?

Thanks all,
Ben

Thanks. I'm really glad I upgraded from a XP3200+.
I'm using the the Cooler Master Hyper 6 Heatsink with 2 Panaflow L1A fans
Unfortunately MSI is having problems with the temp readings on thier A64 boards (especially w/ newcastle cores) so I have no idea what my actual temps are at the moment but the difference between idle and full load is about 5c in my Sonata case. The Hyper 6 is a monster air cooler.

Lighted Picture
UV Picture
CPU-ID

 

[oldfart]

Originally posted by: Markfw900

Originally posted by: oldfart

Originally posted by: Markfw900
OK, no more antagonism. But synthetic benches are crap. Just ask for something I can run, and I will try. But really, flames aside, Intel is hurting right now, and can't compete in anything most people care about. I can set the computer to encode overnight, and the 5 minutes it will save on a 2 hour home movie don't impress me.

Give be something real to run and I will.....

5 mins on a 2 hr home movie? Its more than that. What software are you using? The two I mainly use are Pinnacle Studio 9 and TMPGEnc.
THG benches (I know, I dont like Toms either, but he does use these apps for benches).

P4 3.2 is ~ 20% faster the A64 3200+ in Studio 9 and TMPGEnc. That is WAY more than 5 min on a 2 hr movie.

For games in that same article
Wolf ET, both ~ the same (148 Vs 150 FPS)
Serious Sam again, 101 Vs 106 FPS
Comanche 4 66 Vs 68 no difference
UT2K3, A64 wins here 278 Vs 249 about a 13% advantage
Splinter Cell A64 also wins 85 Vs 71
X2 A64 157 Vs 148
Aquamark same performance

For gaming they both have very playable frame rates in every bench. Many were ~ a tie. In none did the A64 have a 20% advantage like the P4 has in encoding. It depends on what is important to you.

For me, 101 Vs 106 or 278 Vs 249 FPS doesn't matter. You cant tell the difference. 6 Vs 7 Hr encode time is not all that big a deal either.

I think more is made of one Vs the others advantage or disadvantage than is reality. Either one is very fast no matter how you slice it.

Since I don't like to argue small points, I actually have to agree. There is not that much difference, except the Athlon64 wins more in games (real-time, counts more by me) and it runs cooler and cost less (last I looked). So, all in all, I get the Athlon64.

Not looking to argue either. I just find it odd that I see so many posts of how the P4 "sucks" for gaming like its a Celeron or something, or the A64 crushes the P4 in games. Then I go look at some benches to see they are not that far apart. Pricing (using 3.2 as an example) is ~ the same. I do like the lower power requirement and cooler running of the A64.