Affect HD performance has on 3dmark scores ?

[Major]

Just curious, if defraging hd increases 3dmark scores then just how important is hd performance ? anyone know ?

[saaya]

not really important... defraging might give you 50points, maybe 200 if your drive is really fragmented very bad...

its not worth to get a faster hd because of gaming performence...

[q149]

I also wondered this .. if i make a benching partion i get horrible sandra scores on it for some reason like 21k.. and normal drive i get about 38k.

[saaya]

sandras hd bench is weird... i get way better scores with a fragmented crapped drive than with a frsh defraged clean drive...

[antipop]

Use hd tach to mesure hdd performance, it's way more reliable than sandra

[CCW]

I dont see how HDD performance can affect 3DMark much.

Craig

[Hell-Fire]

What about a 10-15k rpm Raid 0 (maybe even SCSI) versus standard 7200 ide?

[mdzcpa]

When I swapped from two PATA ATA 133 Maxtors in RAID 0 to two SATA Raptors in RAID 0, I gained nothing in 3Dmark.

Of course with PCMark, Winbench, and general day to day usage it made a big difference. I was actually stunned with just how much seat of the pants improvement there was.

[Ace-a-Rue]

Originally posted by mdzcpa
When I swapped from two PATA ATA 133 Maxtors in RAID 0 to two SATA Raptors in RAID 0, I gained nothing in 3Dmark.

Of course with PCMark, Winbench, and general day to day usage it made a big difference. I was actually stunned with just how much seat of the pants improvement there was.

What was the model of the Maxtor drive, D740X?

[mdzcpa]

Originally posted by Ace-a-Rue
What was the model of the Maxtor drive, D740X?

Yes sir. They were a pair of D740X-6L 40G drives.

[Ace-a-Rue]

Originally posted by mdzcpa
Yes sir. They were a pair of D740X-6L 40G drives.

I can't remember but does those Raptors come with 8mb of cache?

[saaya]

8mb cache 10Krpm...

[Ace-a-Rue]

So, they can kick some booty!:D

[mdzcpa]

Originally posted by Ace-a-Rue
So, they can kick some booty!:D

Yes they do indeed:
http://www.liquidninjas.com/mdzcpa/forumpics/raptors/atto1.jpg

[Ace-a-Rue]

Very nice!:)

[GVCryan]

I have a Seagate 60g 5400 rpm and a set of 2 wd800jb 8mb drives.

Back when I ran raid0 withe the WD drives, I actually was slower in 3dmark than on the 5400rpm. A single 80 was the best.

I think that whichever system of harddrives have a lower CPU utilization is the superior 3dmark drive, which is why I feel sata will not become popular with 3dmark benchers until some quality chips come out.

Just buy a nice cheap drive for your 3dmark.

[mdzcpa]

Originally posted by GVCryan
Just buy a nice cheap drive for your 3dmark.

Exactly! I have a small 20G WD 7200 ATA 100 that I use only for 3Dmark. All that's on there is the OS (tweaked for benching), some basic utils, and 3Dmark.

[stergiopilus]

HDD speed makes absolutely no difference for my 3dmark2001 scores. I tried running 3dmark out of my ram (ramdrive) and there was no improvements in the score over just running it off the hard drive.

[Ewok]

Originally posted by GVCryan
I have a Seagate 60g 5400 rpm and a set of 2 wd800jb 8mb drives.

Back when I ran raid0 withe the WD drives, I actually was slower in 3dmark than on the 5400rpm. A single 80 was the best.

I think that whichever system of harddrives have a lower CPU utilization is the superior 3dmark drive, which is why I feel sata will not become popular with 3dmark benchers until some quality chips come out.

Just buy a nice cheap drive for your 3dmark.

but cpu speed doesnt affect 3dmark 2003 scores so that doesnt even matter :P

[Ace-a-Rue]

Originally posted by Ewok
but cpu speed doesnt affect 3dmark 2003 scores so that doesnt even matter :P

I disagree with that premise. The CPU maybe has LESS effect in 3d2003 than it did in 3d2001, but the cpu never the less impacts the overall performance of almost any game. I read that the VC can sluff off some chores to the cpu.

I got 5928 at default FSB (200 FSB) and 6048 at 234 FSB. All other bios settings remained the same when I lowered the FSB back to default.

[saaya]

well higher fsb lets the videocard show off some more, but cpu speed really has almost no effect on 2k3...

something a lot of people criticize on 2k3 since all games show a linear performence increase with rising cpu speed...

[Ace-a-Rue]

Grant you its a small change about 2%, but 3d2001 gives me a 8% change with that same fsb. Compare the scores between 2d2001 and 3d2003 it is over 3X difference, or, a better description is 1/3 of what you use to get with 3d2001.

Anyway you look at it, an increase in cpu horsepower will boost your performance.:)

[Dissolved]

my scsi has 0% CPU utilization :)

[Ace-a-Rue]

Originally posted by Dissolved
my scsi has 0% CPU utilization :)

You scsi guys are so incorrigible! hehehehehe:D

[Dissolved]

well scsi doesnt cost much.

the scsi card was $45.00
scsi hdd can run $100~$400 Depending on speed and size
cables are about $10~$35
term is about $10.00

so about 400-500 for a good scsi os drive setup. im not sure how compareable it is to ide or sata raid. but there 0% CPU utilization on my setup as said from hd tach, and the acess times alone for a single scsi drive are about 3.6~5ms if you get a newer drive.

But a pair or raptors would be the better choice, but i wont use one till i get a board with built in sata that Doesnt run off my pci bus :(

[antipop]

Originally posted by Dissolved
well scsi doesnt cost much.

so about 400-500 for a good scsi os drive setup.
:confused: :confused: :confused:
is it only me or 400$ is expensive for a hdd system?

[Dissolved]

hdd system? lol that sounds funny..

but 2-3 ide raid would cost about the same also.

[Ace-a-Rue]

Originally posted by Dissolved
hdd system? lol that sounds funny..

but 2-3 ide raid would cost about the same also.

3 drives would but not 2.

As for % cpu utilization I have about 4% using a 16K stripe. If I jump to 32K stripe, it would be under 2%....

[Dissolved]

not bad. but id rather pay more for scsi anyways. For me its better in every way but price.

Someday when i get a good enough pc and learn not to be lazy ill do More testig and see differences if any between the wd 360 sata 36gb 10k, and a seagate 15k.3 36gb.
ill probly have both soon. so maybe ill run some testing and see if theres a Real difference.

[Ace-a-Rue]

I'd be interested to hear about the results.

[antipop]

Originally posted by Dissolved
hdd system? lol that sounds funny..

but 2-3 ide raid would cost about the same also. I wanted to say hard drive but it's more than just a hard drive because there's also the card. I don't think you can beat ide in the price, you can find 120GB drive for around $120 and there's now 8Mb cache version

[Dissolved]

well you get what you pay for. thats bascially all i can say.

ide and sata are getting alot better but i just dont have the case space for a billion ide drives in raid. but a few scsi drives :)

[Ace-a-Rue]

Sounds like to me justification for scsi...;)

[Dissolved]

SCSI or IDE?

*speed or space?

*2-3 drives or 4-8 drives?

*3-5years or 1-3years?


Thats how i see it :)