*** Official ASUS P4C800/Deluxe (875P) Thread ***

[Chad]

Computer, here is a screenshot of my results...

Sandra Results Pic

 

[computer]

I haven't had time to catch on posts here today or last night. I noticed on the WD740 there is a jumper for PM2 Enabled. I see that's power management. But did any of you know the default jumper position is for spread spectrum? Default is on the 1 & 2 pins which is DISabled. Removing that jumper enables it. Going by spread spectrum for a CPU, it should be disabled for O'clocking. So, I'm wondering if any of you have compared the differences between the jumper on and the jumper off?

Does anyone know if there are any advantages or is it more desirable to use the SATA power connector over the legacy 4 pin Molex? I've often wondered about that. I was using the 4 pin Molex on my WD360 Raptor.

 

[DaveR]

I do not have the WD but do use the SATA power cable from my Antec case supply.

I wonder, if after all of our "benchmarks", I should leave this on the Promise already!

If it would matter (and I hear it would not) I could start over and use a 32k stripe vs. my current 64k stripe.

P.S. Computer is a 40 degree C running 2.4c temp OK?
Also, is there a new ASUS util? I show my drive at 59gb and not 159gb free...ONLY in this ASUS probe program.

 

[Chad]

HDTach 3.00 Beta

Asus p4C800-E-Deluxe
ICH5 Controller

My Results...

Western Digital 74gig SATA150 Raptor

Short
Random Access: 7.8ms
CPU utilization: 3%
Average read: 64.9mbps
Burst: 130.5mbps

Long
Random Access: 7.8ms
CPU utilization: 2%
Average read: 65.0mbps
Burst: 130.6mbps



Western Digital 120JB Special Edition

Short
Random Access: 13.4ms
CPU utilization: 2%
Average read: 41.1mbps
Burst: 97.4mbps

Long
Random Access: 13.5ms
CPU utilization: 0%
Average read: 41.2mbps
Burst: 97.8mbps

 

[DaveR]

OK, I d/l hdtach 3.0 too.

Here is what I got with my 2 Maxtor SATA 80GB RAID 0's on Promise...and a real world eval!

Short
Random Access: 15.5 ms
CPU utilization: 4%
Average read: 92.2 mbps
Burst: 115.9 mbps

Long
Random Access: 15.6ms
CPU utilization: 1%
Average read: 91.7 mbps
Burst: 119.2mbps


So, what does this system I built do, besides run benchmarks?

A bunch of us at work use my VERY large Spreadsheets and presentions. They complain it takes 10 minutes to open...no lie!

With this new system I built it takes me 25 seconds!!!!
Maybe I should stop these benchmarks and leave the system alone!

 

[Chad]

Originally posted by: DaveR
Average read: 92.2 mbps

Veeerry NICE!

 

[DaveR]

...but I keep wondering about the ICH5R. I guess if I do not just rebuild this I will always wonder if the Promise was the best choice?

Originally posted by: Chad

Originally posted by: DaveR
Average read: 92.2 mbps

Veeerry NICE!

:

 

[Chad]

Well, on my system, the ICH5 was better in every bench. I would go with it because it also aleviates and bypasses some stuff if I recall correctly, something to do with a bus of some sort. It's worth switching to. And you don't need to reinstall anything. I switched back and forth between the controllers with ease with absolutely no problems. There's no reason at all that should be a problem.

 

[computer]

Originally posted by: Chad
Well, on my system, the ICH5 was better in every bench. I would go with it because it also aleviates and bypasses some stuff if I recall correctly, something to do with a bus of some sort. It's worth switching to. And you don't need to reinstall anything. I switched back and forth between the controllers with ease with absolutely no problems. There's no reason at all that should be a problem.

Did you run PCmark, Winbench, AIDA32, ATTO and HDtach?

 

[computer]

Originally posted by: senior guy
Hey guys, I need some memory advice:

I'm giving my current comp to my son and will be building another one using the P4C800 Deluxe. I wanted to check here before buying a pair of 512mb dual-channel memory sticks... Which ones would deliver the best performance in this board bearing in mind that I'll be running it stock ?

I have another related question... Given that my first memory pair will be 2 x 512mb, is it ok to put in a second pair at 2x 256mb? ...is that is doable? and if so, would doing that adversely affect performance?

Thanks in advance.

Can't give any accurate advice without knowing your CPU. PC3700 is cutting it pretty close to get DDR500 from it. Some may O'clock to that speed (which is 250mhz bus) and some may not. The slower the CPU, the faster the memory has to be due to the multiplier. For a 2.4C, you'd want at least DDR500/PC4000 which @default would give you 3ghz. For a 3ghz CPU, PC4000 is a waste of money, plus will have rather bad performance since you'd be UNDERclocking it. You'd have to get 3750mhz from a 3ghz P4 to reach the DDR500 spec.

 

[computer]

Originally posted by: IntelConvert
I have a similar question to the 2nd part of senior guy's post in that I now have 2 x 256MB sticks and I now want to add more memory (2 x 512MB as a 2nd pair). I know that my memory timings will probably need to be relaxed somewhat when using 4 DIMMs, but I really don't care so long as it won't result in instability, or lousy performance!



So I too could use some advice here - must all 4 DIMMs be the same brand and model???



------

Edit: Dave, what are KOMUSA sticks?

Always remember that the more sticks you add, the slower your FSB will probably have to be. So, don't expect to get as good of O'clocking results. You'll also probably have to relax timings a bit. And most certainly the memory has to match, it would be even much worse if the sticks were not the same. Until someone makes a 4 stick dual channel kit, 4 sticks is going to cause some problems. And yes, you could also be looking at some instability. You need to at least get the EXACT same sticks; same brand, same size, same chips, etc. for the best results.

 

[DaveR]

But, I would need the Intel drivers wouldn't I? The ones you install using F6?

Perhaps I can just add new hardware and plug in the floppy?

Originally posted by: Chad
Well, on my system, the ICH5 was better in every bench. I would go with it because it also aleviates and bypasses some stuff if I recall correctly, something to do with a bus of some sort. It's worth switching to. And you don't need to reinstall anything. I switched back and forth between the controllers with ease with absolutely no problems. There's no reason at all that should be a problem.

 

[DaveR]

Computer...you have to admit that my pc4200 Komusa memory has been doing very well. 275MHZ FSB (so far) at 1:1.

Originally posted by: computer

Originally posted by: senior guy
Hey guys, I need some memory advice:

I'm giving my current comp to my son and will be building another one using the P4C800 Deluxe. I wanted to check here before buying a pair of 512mb dual-channel memory sticks... Which ones would deliver the best performance in this board bearing in mind that I'll be running it stock ?

I have another related question... Given that my first memory pair will be 2 x 512mb, is it ok to put in a second pair at 2x 256mb? ...is that is doable? and if so, would doing that adversely affect performance?

Thanks in advance.

Can't give any accurate advice without knowing your CPU. PC3700 is cutting it pretty close to get DDR500 from it. Some may O'clock to that speed (which is 250mhz bus) and some may not. The slower the CPU, the faster the memory has to be due to the multiplier. For a 2.4C, you'd want at least DDR500/PC4000 which @default would give you 3ghz. For a 3ghz CPU, PC4000 is a waste of money, plus will have rather bad performance since you'd be UNDERclocking it. You'd have to get 3750mhz from a 3ghz P4 to reach the DDR500 spec.

 

[computer]

Dave and Chad...and whomever else is running the bench's, remember I only tested single drives, I didn't do any RAID tests. The ICH5 may be faster than the Promise in RAID, but it certainly is not in a single drive setup in all the bench's I ran. You could see the Sandra '04 results. It's a different version, but what I don't understand is the version I have is newer that what you guys are using. 2004.2. comes before 2004.10., doesn't it? Remember that Sandra sucks as a real benchmark tool. It is wildly inconsistent. It has to be run repeatedly then you have to get the averages from all the runs and use that as the results. I've noticed some amazing differences between versions of Sandra in the past. I think it was v98 or something back then, and going from it to v2000 or 2001 had results not anywhere close to each other. Then, from v2002 to v2004 there was another massive difference in results. However I've never seen any differences between build versions of the same program version, or any differences between the Pro and free version.

HDTach, AIDA32, PCmark '04, ATTO, and Winstone are the tests to use. You have to get the latest version of AIDA32 since it has the HD tools. The version of HDtach I used was 2.61.

Chad, that's some rather outstanding results for a WD800JB on HDTach for the max read. I never got over 59k in my tests, and like the rest the highest result was on the Promise. Maybe that's due to v3.0 you are using of HDtach. I'm going to finish my tests of the WD740 on the Promise, then test it on the ICH5 so I can add that to my "database", then I'll try HDtach 3.0 and see if my version of Sandra will update, then run those again.

BTW, my WD740 is smoking every other drive's results in all but maybe 2 specific tests. On HDtach I got 116k max read and about 62k average on the Promise. So, that 130k figure someone got is better than mine, maybe that's v3.0 again doing that. The WD360 was the prior winner with 85k on Promise and 77k on the ICH5. Average was about 46-47k on both controllers. So, you can see how much faster the WD740 is over the WD360. However the write portion of HDtach showed disappointing results for the WD740 where it only got 4th place. I don't understand that. Plus there was one area of Winbench where it did worse than the WD360. All of the other areas it smoked it.

You can get Sandra Pro 2004 here., although that appears to be a newer build that what I have. Mine is called fo-ssm4 and that one is fo-ssm41. 41 is also about 1mb larger.

 

[computer]

Originally posted by: DaveR
Computer...you have to admit that my pc4200 Komusa memory has been doing very well. 275MHZ FSB (so far) at 1:1.

Originally posted by: computer

Originally posted by: senior guy
Hey guys, I need some memory advice:

I'm giving my current comp to my son and will be building another one using the P4C800 Deluxe. I wanted to check here before buying a pair of 512mb dual-channel memory sticks... Which ones would deliver the best performance in this board bearing in mind that I'll be running it stock ?

I have another related question... Given that my first memory pair will be 2 x 512mb, is it ok to put in a second pair at 2x 256mb? ...is that is doable? and if so, would doing that adversely affect performance?

Thanks in advance.

Can't give any accurate advice without knowing your CPU. PC3700 is cutting it pretty close to get DDR500 from it. Some may O'clock to that speed (which is 250mhz bus) and some may not. The slower the CPU, the faster the memory has to be due to the multiplier. For a 2.4C, you'd want at least DDR500/PC4000 which @default would give you 3ghz. For a 3ghz CPU, PC4000 is a waste of money, plus will have rather bad performance since you'd be UNDERclocking it. You'd have to get 3750mhz from a 3ghz P4 to reach the DDR500 spec.

Sure. You have a 2.4C though, right? You wouldn't be getting those speeds with a 3.0ghz CPU and maybe not even a 2.8.

 

[computer]

Thanks for helping us out here. I hope we can figure out what is wrong as I really want to get the most out of this hard drive. So far I have been very unimpressed. My Western Digital 120jb Special Edition is almost this drives equal, and is bigger MUCH MUCH quiter (dang this Rapor is really loud - my 120jb is dead silent!).

WOW, I don't see how that's possible. In all my tests the WD800JB was among the slowest and the WD360 old Raptor smoked it in all but one test. My WD360 would whine a bit, but that's to be expected from a 10k RPM drive, but the WD740 is much quieter than the WD360.

Dave I don't recall where to get the benchmark programs, but any search will turn them up. They are all free I believe. I think I had to find a cr*ck for one of them though. I'll have to check which one.

Ok I'm off ti finish my tests now. Then I'll try the new versions of HDTach and that version of Sandra at that link I just posted.

 

[DaveR]

Yup, 2.4c with the good stepping.

Originally posted by: computer

Originally posted by: DaveR
Computer...you have to admit that my pc4200 Komusa memory has been doing very well. 275MHZ FSB (so far) at 1:1.

Originally posted by: computer

Originally posted by: senior guy
Hey guys, I need some memory advice:

I'm giving my current comp to my son and will be building another one using the P4C800 Deluxe. I wanted to check here before buying a pair of 512mb dual-channel memory sticks... Which ones would deliver the best performance in this board bearing in mind that I'll be running it stock ?

I have another related question... Given that my first memory pair will be 2 x 512mb, is it ok to put in a second pair at 2x 256mb? ...is that is doable? and if so, would doing that adversely affect performance?

Thanks in advance.

Can't give any accurate advice without knowing your CPU. PC3700 is cutting it pretty close to get DDR500 from it. Some may O'clock to that speed (which is 250mhz bus) and some may not. The slower the CPU, the faster the memory has to be due to the multiplier. For a 2.4C, you'd want at least DDR500/PC4000 which @default would give you 3ghz. For a 3ghz CPU, PC4000 is a waste of money, plus will have rather bad performance since you'd be UNDERclocking it. You'd have to get 3750mhz from a 3ghz P4 to reach the DDR500 spec.

Sure. You have a 2.4C though, right? You wouldn't be getting those speeds with a 3.0ghz CPU and maybe not even a 2.8.

 

[DaveR]

I did get all the Benchmark programs Ok. I also signed up for, and got hdtach 3.

Originally posted by: computer

Thanks for helping us out here. I hope we can figure out what is wrong as I really want to get the most out of this hard drive. So far I have been very unimpressed. My Western Digital 120jb Special Edition is almost this drives equal, and is bigger MUCH MUCH quiter (dang this Rapor is really loud - my 120jb is dead silent!).

WOW, I don't see how that's possible. In all my tests the WD800JB was among the slowest and the WD360 old Raptor smoked it in all but one test. My WD360 would whine a bit, but that's to be expected from a 10k RPM drive, but the WD740 is much quieter than the WD360.

Dave I don't recall where to get the benchmark programs, but any search will turn them up. They are all free I believe. I think I had to find a cr*ck for one of them though. I'll have to check which one.

Ok I'm off ti finish my tests now. Then I'll try the new versions of HDTach and that version of Sandra at that link I just posted.

 

[DaveR]

I did put my 2 drives on the ICH5R, but I see no way to set them up without killing what is on there now. Yes, I can put a small hard drive on the PATA port, and run Drive Image, etc, but short of that I do not see how I can switch from Promise to Intel just by moving the cables.

Am I missing something here?

 

[computer]

I did get all the Benchmark programs Ok. I also signed up for, and got hdtach 3.

HDtach 3 is beta, so I wouldn't put too much weight on those results but I may try it anyway. How did you d'load it? Is it a generic link or a proprietary link for your email address? If it's a general link, please post it since I'm not about to give any of my email addresses out for something like that. I don't trust ANY place regardless of how "safe" they say your email address may be.

 

[computer]

Originally posted by: computer

I did get all the Benchmark programs Ok. I also signed up for, and got hdtach 3.

HDtach 3 is beta, so I wouldn't put too much weight on those results but I may try it anyway. How did you d'load it? Is it a generic link or a proprietary link for your email address? If it's a general link, please post it since I'm not about to give any of my email addresses out for something like that. I don't trust ANY place regardless of how "safe" they say your email address may be.

Never mind, I found the actual link.

 

[Chad]

DaveR, yeah you're right, I'm sorry. I have a single non raid drive, so I was speaking to that. I don't know if it's possible to swtich a raid drive's controller without losing data (I would assume you cant).

Computer, I didn't mean my 120jb was as fast as the Raptor per se, and if you look at my post above with the HDTach bench of both drives, you will see it's not even close. But in just normal everyday use, I can't really percieve the difference while milling around in windows. Also, the 120jb is dead silent makes absolute 0, nala, NO noise at all, while I can hear the Raptor heads clanging around a lot when it seeks and stuff.

Also, in this thread, most of the people in there that have Raptors are scoring almost identical scores as mine (you thought my 130 was impressive - but you'll see it's common for this drive actually).

I will have to bench the Raptor on the Promise controller with HDTach 3.0 this weekend to compare it with the ICH5 scores I posted above.

 

[computer]

How did I subscribed to that thread? I just got an email notification on it. I'll try HDtach 3 or 2.9 whatever it is shortly. It looks completely different and also has different results I see.

 

[computer]

This is pretty pathetic!! Only 98.7 burst, 64.8 average for the WD740. Access was 7.8ms, CPU 3%.

Long test was even worse at 95 burst but average was 65.1, 7.7ms and 4%.

 

[computer]

This is impossible. I just ran it again, and accidentally on the OS drive, and old IBM 20gb 7200rpm w/only 2mb buffer, ATA100, and got 101 burst!! HIGHER than the friggin' WD740!!! Beta crap no doubt. Oh yeah, this sucks. I just ran it again on the WD740, and get this: Short burst 1103.5mb/sec!! That's from about 98mb/sec on the first try, then 1103mb/sec on the 2nd try!! Then the long test was back down to 98!! This is about as reliable as a crack whore.