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

[DaveR]

computer what Sandra are you using? You quote data 10x "better" and drive index Chad and I see is in MB/sec and not kb/sec.

 

[computer]

This isn't going to be laid out correctly and it may be hard to read since this form is HTML enabled, and all of my formatting on this is going to be lost. This is just a paste from Wordpad. But, you should be able to figure it out. It would take too long for me to put in all of the / marks I put on the Maxtor results, so just note that how the Maxtor results are laid out with the /, that should go where there are spaces bettween the number results for the other drives. I haven't finished this documentation yet, so there may be some things left out or comments not complete.

SiSoft Sandra Pro 2004

These results are with Windows Cache used ("Bypass Windows cache"
box is UNchecked, NOT checked).

Blue = Best scores of all drives in this test.
Red = Best scores for a given drive comparing controllers, except if Blue
which is better.



Maxtor:

Native IDE / Promise

Index: 38315 / 36681
Buffered Read: 66 / 1476
Sequential Read: 57 / 57
Random Read: 11 / 11
Buffered Write: 75 / 1357
Sequential Write: 42 / 32
Random Write: 18 (tie) / 13
Avg. Access Time ms: 5 / 5

-------------------------------------------------------------


Raptor:

Native ICH5 Promise

Index: 30921 31264
Buffered Read: 98 1465
Sequential Read: 42 45
Random Read: 14 15
Buffered Write: 73 1514
Sequential Write: 37 29
Random Write: 18 (tie) 17
Avg. Access Time ms: 3 3

---------------------------------------------------------------


Seagate 120:

Native IDE Promise

Index: 34563 33658
Buffered Read: 83 961
Sequential Read: 53 53
Random Read: 12 12
Buffered Write: 72 1084
Sequential Write: 29 23
Random Write: 14 13
Avg. Access Time ms: 4 4

--------------------------------------------------------------


Seagate 80:

Native IDE Promise

Index: 35128 32712
Buffered Read: 86 1456
Sequential Read: 51 49
Random Read: 11 11
Buffered Write: 71 1388
Sequential Write: 46 31
Random Write: 13 13
Avg. Access Time ms: 4 4

--------------------------------------------------------------


WD800JB:

Native IDE Promise

Index: 35157 34273
Buffered Read: 84 952
Sequential Read: 53 53
Random Read: 12 12
Buffered Write: 77 1064
Sequential Write: 30 23
Random Write: 17 16
Avg. Access Time ms: 4 4

---------------------------------------------------------------



Raptor on Si3112a PCI SATA controller card:
Results for the Raptor on the PCI controller card are given as:
With integrated P4C800-E Deluxe Promise controller: ENabled/DISabled.
One score indicates same results.

Bold = Best score comparing the integrated Promise controller enabled/disabled.
Green = Best score for the Raptor on all controllers.
Red = Better than Raptor on ICH5

Index: 29559/28636
Buffered Read: 97/96
Sequential Read: 39/38
Random Read: 12
Buffered Write: 83/39
Sequential Write: 40/38
Random Write: 17
Avg. Access Time ms: 4

Note: Note how overall results are a bit worse when the integrated Promise controller
of the P4C800-E Deluxe was DISabled even though it was not being used! Also, note
how the Si3112a PCI controller card had the poorest performance of all controllers,
(especially with none of the "astronomical" buffered read and buffered write results
that is seen on the integrated Promise controller), except for the Buffered Write
rates where it had the highest compared to ALL of the the tested hard drives'
Native controllers, and the Sequential Write rates which were the highest for the
Raptor period, but not by much. Overall, IMO for the Raptor as far as Sandra is
concerned, the integrated P4C800-E Deluxe Promise controller still wins UNLESS
Sequential Write rates are more important to one. The integrated Promise
controller of the P4C800-E Deluxe seems to be "lacking" in the area of Sequential
Write rates showing the worst performance of ANY of the controllers regardless of
drive tested, again, only as far as Sequential Writes go.


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


Individual Results Standings:
(In MB/sec except where noted. Higher is better, except
"Access Time" where lower is better).


Index: (in KB/sec)

1. 38,315 Maxtor on IDE
2. 36,681 Maxtor on Promise
3. 35,157 WD800 on IDE
4. 35,128 Seagate 80 on IDE
5. 34,563 Seagate 120 on IDE
6. 34,273 WD800 on Promise
7. 33,658 Seagate 120 on Promise
8. 32,712 Seagate 80 on Promise
9. 31,264 Raptor on Promise
10. 30,921 Raptor on ICH5R
11. 29,559 Raptor on Si3112a



Buffered Read:

1. 1476 Maxtor on Promise
2. 1465 Raptor on Promise
3. 1456 Seagate 80 on Promise
4. 961 Seagate 120 on Promise
5. 952 WD800 on Promise
6. 98 Raptor on ICH5R
7. 97 Raptor on Si3112a
8. 86 Seagate 80 on IDE
9. 84 WD800 on IDE
10. 83 Seagate 120 on IDE
11. 66 Maxtor on IDE



Sequential Read:

1. 57 TIE; Maxtor on IDE & Maxtor on Promise
3. 53 TIE; WD800 on IDE, WD800 on Promise,
Seagate 120 on Promise & Seagate 120 on IDE
7. 51 Seagate 80 on IDE
8. 49 Seagate 80 on Promise
9. 45 Raptor on Promise
10. 42 Raptor on ICH5R
11. 39 Raptor on Si3112a



Random Read:

1. 15 Raptor on Promise
2. 14 Raptor on ICH5R
3. 12 TIE; Seagate 120 on Promise, Seagate 120 on IDE,
WD800 on IDE, WD800 on Promise & Raptor on Si3112a
11. 11 TIE; Seagate 80 on IDE, Seagate 80 on Promise,
Maxtor on IDE & Maxtor on Promise



Buffered Write:

1. 1514 Raptor on Promise
2. 1388 Seagate 80 on Promise
3. 1357 Maxtor on Promise
4. 1084 Seagate 120 on Promise
5. 1064 WD800 on Promise
6. 83 Raptor on Si3112a
7. 77 WD800 on IDE
8. 75 Maxtor on IDE
9. 73 Raptor on ICH5R
10. 72 Seagate 120 on IDE
11. 71 Seagate 80 on IDE



Sequential Write:

1. 46 Seagate 80 on IDE
2. 42 Maxtor on IDE
3. 40 Raptor on Si3112a
4. 37 Raptor on ICH5R
5. 32 Maxtor on Promise
6. 31 Seagate 80 on Promise
7. 30 WD800 on IDE
8. 29 TIE; Raptor on Promise & Seagate 120 on IDE
11. 23 TIE; WD800 on Promise & Seagate 120 on Promise



Random Write:

1. 18 TIE; Maxtor on IDE & Raptor on ICH5
3. 17 TIE; Raptor on Si3112a, WD800 on IDE,
& Raptor on Promise
6. 16 WD800 on Promise
7. 14 Seagate 120 on IDE
11. 13 TIE; Seagate 80 on Promise, Seagate 80 on IDE,
Seagate 120 on Promise & Maxtor on Promise



Average Access Time (ms):

1. 3 TIE; Raptor on ICH5 & Raptor on Promise
3. 4 TIE; Seagate 80 on Promise, Seagate 80 on IDE,
WD800 on Promise, WD800 on IDE, Raptor on Si3112a,
Seagate 120 on IDE & Seagate 120 on Promise
11. 5 TIE; Maxtor on Promise & Maxtor on IDE


Maxtor/Raptor tie as winner on Native controller. Not counting access,
Maxtor is winner. It would of course depend on the user as to which of
these test results are more important.

Raptor is winner on Native controller for number of tests won. Not
counting access time, Raptor & Maxtor tie. It would of course depend
on the user as to which of these test results are more important.

As for the overall controller winner; [comment] all results are so close EXCEPT
FOR the Buffered Read/Buffered Write results where the Promise controller
blows away the Native controller, that I'd have to say the Promise controller is
the overall winner.

 

[Chad]

Yeah Dave, yours look about in line with my results with of course yours being higher because you have Raid0 and I do not (I have a single drive) and also you run a much higher FSB, so that is expected. That accounts for your better scores.

Comparing you results to mine, it is obvious that whatever we are doing, we are jiving. Your scores are above mine about right where they should be considering your Raid0 and FSB. I am starting to think computer is running an old version of Sandra with his kb/s comment (?). There is something fundamentally different here I think.

Also, are your Raptors loud when the seek? Mine makes a lot of noise when it seeks hard or especially if i defrag it. My Western Digital 120jb Special Edition drive is dead quiet so I'm not used to this... is this normal?

 

[computer]

Originally posted by: DaveR
computer what Sandra are you using? You quote data 10x "better" and drive index Chad and I see is in MB/sec and not kb/sec.

I mentioned before it's Sandra Professional 2004 (Unicode Win32 x86, 2004.10.9.89). All of the drive indexes are in KB/sec and it's been this way on every version since '98. So I don't know why some are seeing that number in MB/sec. This the number that is superimposed OVER the colored bar graph at the top to the right of "Current drive:"

 

[computer]

This is the details for the test setup. Again, formatting will be lost so it may be hard to read. This will also explain the details of the drives tested.



Test platform for hard drive tests:
(1-2004)


Motherboard: Asus P4C800-E Deluxe (BIOS v1014) with motherboard
drivers installed from the CD. LAN was enabled, IEEE1394
was disabled, audio was enabled. "Legacy USB" was disabled.
"PAT" enabled, "Turbo" mode was enabled.

CPU: P4 2.4C (HyperThreaded, 800mhz bus) running @ about
3.4ghz with a 281mhz bus.

Memory: Geil Ultra Platinum PC4000 @2.5-4-4-7 running 1:1 ratio
for DDR562 running in dual channel mode.

Video: "Cheap" Nvidia GeForce2 MX 4x AGP 64mb with latest 5053
drivers installed.

Power supply: Fortron-Source FSP400-60PFN

OS: Windows XP Pro w/SP1 (NOT SP1a)

Notes: No cards/peripherals were installed (except of course for a
the video card). All hard drives were blank, just formatted
drives using their respective manufacturer's format CD in
NTFS. All drives were stand alone master configurations.
No case was used. Main OS hard drive was an IBM 60GXP
20gb, 7200rpm, 2mb buffer. Drives were frequently
defrag'd during the tests.


Details for hard drives:


"Maxtor" = DiamondMax Plus 9 6Y080P0 80gb, 7200rpm, 8mb cache,
ATA133/UDMA 6.

"Raptor" = Western Digital original 34.7gb Raptor, WD360GD, 10,000rpm,
8mb cache, SATA150 (Serial ATA) UDMA 6.

"Seagate 120" = ST3120026A Barracuda 120gb, 7200rpm, 8mb cache, ATA100,
UDMA 5.

"Seagate 80" = ST380013A Barracuda 80gb, 7200rpm, 8mb cache, ATA100,
UDMA 5.

"WD800/JB" = Western Digital WD800JB Caviar Special Edition 80gb, 7200rpm,
8mb cache, ATA100/UDMA 5.

"Native" = Refers to the Native controllers on the motherboard. This would
be the typical IDE controller/connectors for the IDE drives, and
the ICH5(R) which is the Native controller for the serial ATA Raptor.

"IDE" = Same as above in the case of the IDE drives. The words "IDE" and
"Native" will be used interchangeably in the results.

"Promise" = Refers to the integrated Promise PDC20378 controller chip on this
motherboard. This controller has an IDE connector for IDE type
drives, and 2 SATA150 serial ATA connectors for serial ATA
drives. (2nd is for RAID configurations which was not tested).


Notes:

It should be noted that is this BIOS revision, the Native controllers (ICH5 &
IDE) are limited to only UDMA 5, (thank Asus for that) and the Raptor &
Maxtor are UDMA 6 drives. This probably accounts for some of the Raptor's
poor showing on some tests on the ICH5 controller. However, this did not
seem to affect the Maxtor as much when it was on the Native controller.
Probably due to the Maxtor being ATA133, and the Raptor being sATA150;
there's more of a transfer protocol difference there. The Promise controller
on the other hand can support UDMA 6 in this BIOS revision, which is
probably partially to thank for some of its performance in addition to the fact
this controller also uses more buffering/caching than Native controllers.

 

[Chad]

Originally posted by: computer

Originally posted by: DaveR
computer what Sandra are you using? You quote data 10x "better" and drive index Chad and I see is in MB/sec and not kb/sec.

I mentioned before it's Sandra Professional 2004 (Unicode Win32 x86, 2004.10.9.89). All of the drive indexes are in KB/sec and it's been this way on every version since '98. So I don't know why some are seeing that number in MB/sec. This the number that is superimposed OVER the colored bar graph at the top to the right of "Current drive:"

Yup, different Sandra, mine is not Pro and is free. It is Sisoft Sandra Standard 2004 (Unicode Win32 x86, 2004.2.9.104) so that perhaps accounts for the difference (?).

 

[DaveR]

Well, with bypass unchecked I get this!

Promise
File System Drive Index - 29MB/s
Buffered Read - 47MB/s
Sequential Read - 31MB/s
Random Read - 14MB/s
Buffered Write - 87MB/s
Sequential Write - 47MB/s
Random Write - 41MB/s
Average Access Time - 38ms


I sure do not know how you are getting the numbers you are. These are NOT blank drives, but my OS/Data drives.

 

[computer]

Originally posted by: Chad

Originally posted by: computer

Originally posted by: DaveR
computer what Sandra are you using? You quote data 10x "better" and drive index Chad and I see is in MB/sec and not kb/sec.

I mentioned before it's Sandra Professional 2004 (Unicode Win32 x86, 2004.10.9.89). All of the drive indexes are in KB/sec and it's been this way on every version since '98. So I don't know why some are seeing that number in MB/sec. This the number that is superimposed OVER the colored bar graph at the top to the right of "Current drive:"

Yup, different Sandra, mine is not Pro and is free. It is Sisoft Sandra Standard 2004 (Unicode Win32 x86, 2004.2.9.104) so that perhaps accounts for the difference (?).

Naaa, I don't think. I got the same results with the non-pro version. This is very strange.

 

[DaveR]

Mine is Pro and is 2004.2.9.104

Originally posted by: Chad

Originally posted by: computer

Originally posted by: DaveR
computer what Sandra are you using? You quote data 10x "better" and drive index Chad and I see is in MB/sec and not kb/sec.

I mentioned before it's Sandra Professional 2004 (Unicode Win32 x86, 2004.10.9.89). All of the drive indexes are in KB/sec and it's been this way on every version since '98. So I don't know why some are seeing that number in MB/sec. This the number that is superimposed OVER the colored bar graph at the top to the right of "Current drive:"

Yup, different Sandra, mine is not Pro and is free. It is Sisoft Sandra Standard 2004 (Unicode Win32 x86, 2004.2.9.104) so that perhaps accounts for the difference (?).

 

[Chad]

With bypass unchecked and everything else checked, this is what I get...

ICH5
File System Drive Index - 42MB/s
Buffered Read - 83MB/s
Sequential Read - 41MB/s
Random Read - 53MB/s
Buffered Write - 73MB/s
Sequential Write - 31MB/s
Random Write - 27MB/s
Average Access Time - 6ms (estimated)

 

[computer]

Originally posted by: DaveR
Well, with bypass unchecked I get this!

Promise
File System Drive Index - 29MB/s
Buffered Read - 47MB/s
Sequential Read - 31MB/s
Random Read - 14MB/s
Buffered Write - 87MB/s
Sequential Write - 47MB/s
Random Write - 41MB/s
Average Access Time - 38ms


I sure do not know how you are getting the numbers you are. These are NOT blank drives, but my OS/Data drives.

That's pretty hideous results except for Buffered Write. You're supposed to get better results with cache bypassed since that will test raw controller performance. Blank drives are going to be a bit different, but not by that much. I was going to upload a screenshot of Sandra, but I can't login to my website right now. I opened a support ticket and maybe they will have it fixed soon.

 

[computer]

Maybe you all should try PCmark '04, AIDA32 (latest version w/disk tests plugin) or Winbench. They are much more accurate and MUCH more consistent than Sandra.

I have to get back to work, but when I'm able to login to my site I'll upload a screenshot so you can see what mine looks like.

 

[Chad]

email me the pic at chadsparks@cox.net and i'll host it for ya

also, computer, I think you should try doing the update within the program, seriously. i think that's the reason we are getting such different results.

p.s. neither of you answered the question i posed about the drive being so loud. is this normal?

 

[computer]

Mine isn't loud, but I have all sorts of fans running and a huge air cleaner.

Here is the screenshot I uploaded.

 

[Chad]

Is that with those two options you mentioned earlier UNchecked?

 

[computer]

Originally posted by: Chad
Is that with those two options you mentioned earlier UNchecked?

You can see in the screenshot what's enabled and what is disabled. See the red "X"s. The IO didn't matter that much, but bypassing cache does make a pretty big difference. So, at the top of this where you see "37750Kb/sec" yours doesn't have that area?

 

[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.

 

[Thor86]

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.

For best performance, I would go with the OCZ PC3500EBs, and if you plan on overclocking beyond 250fsb, then OCZ PC3700EBs. If you don't care too much about memory timings, then go with generic PC3200s.

 

[DaveR]

Chad, I am using two Maxtor 80GB SATA drives. I did not get the Raptors...but looked at them!

Originally posted by: Chad
Yeah Dave, yours look about in line with my results with of course yours being higher because you have Raid0 and I do not (I have a single drive) and also you run a much higher FSB, so that is expected. That accounts for your better scores.

Comparing you results to mine, it is obvious that whatever we are doing, we are jiving. Your scores are above mine about right where they should be considering your Raid0 and FSB. I am starting to think computer is running an old version of Sandra with his kb/s comment (?). There is something fundamentally different here I think.

Also, are your Raptors loud when the seek? Mine makes a lot of noise when it seeks hard or especially if i defrag it. My Western Digital 120jb Special Edition drive is dead quiet so I'm not used to this... is this normal?

 

[DaveR]

I answered that I don't use Raptors.

Computer, where are those other benchmark programs you mentioned?

Originally posted by: Chad
email me the pic at chadsparks@cox.net and i'll host it for ya

also, computer, I think you should try doing the update within the program, seriously. i think that's the reason we are getting such different results.

p.s. neither of you answered the question i posed about the drive being so loud. is this normal?

 

[DaveR]

Computer, mine has it but in MB/sec.

Originally posted by: computer

Originally posted by: Chad
Is that with those two options you mentioned earlier UNchecked?

You can see in the screenshot what's enabled and what is disabled. See the red "X"s. The IO didn't matter that much, but bypassing cache does make a pretty big difference. So, at the top of this where you see "37750Kb/sec" yours doesn't have that area?

 

[DaveR]

I am having great luck with the KOMUSA sticks. I told them to use AData chips and a "pretty" pair of copper heat spreaders and thermal tape.

The 1Gb (2x512) pair came to $272 and it is PC4200. My FSB is 275Mhz so it is running at 550MHZ now...with a 2.4c Proc.

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.

 

[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?

 

[DaveR]

I went to

KOMUSA

EDIT...I see this is broken, too on the new forum. Just click KOMUSA.

I ordered the hyperam (use that name in coupon code area for free shipping). You will see an area to specify AData chips with Copper spreaders.

Also, call them. They will tell you if you can mix memory.

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?

 

[DaveR]

Computer...dont forget I am also using a 64k stripe. That does slow down benchmarks.