Does anyone know of a motherboard that has all of these qualifications?

Intel 915 or 945 chipset
Supports P4 600 series CPU
Supports SATA 2 300mb/s
Has both PCI-x and AGP

I recently bought a Leadtek A400 GTX AGP card. I want to upgrade to a better motherboard that is faster than my AMD Barton 2600. I want a smoking fast hard disk. I want to be able to use my video card but also should I decide to buy a 7800 I dont want to buy a new motherboard. There are several boards that have the 915 chipset that have agp and pci-x but they only support sata 150. I really want the double speed drive.

Also who knows about scsi? How does a scsi card and drive compare to sata 150 and sata 300 in data transfer, assuming I'm gaming and not running a server. I cant find any real comparison data on tomshardware. Does a scsi320 card on pci-x kick ass on sata 300?

SCSI rocks ATA. A lot.

The way ATA works, your CPU is somewhat heavily involved in the I/O. With SCSI, much less so. So, it's not only faster I/O, but lets the CPU devote more time to other processes.

SCSI is faster, however SATA 2, aka SATA 3G, Aka a couple other monikers is incredibly fast, and backwards compatable should you need. As well, make sure those PCI express are the proper PCI-ex for putting video cards into, there is a difference between them and the PCI-ex 1, and the PCI-ex 16. IE one is really really tiny, the other is aboutthe same size as AGP.

Though I generally don't like intel, both my laptop, and my media center have the 915 chipset with Pentium Ms, and I have to say I am fairly impressed with them.

If you're gaming and not running a very serious server, why spend hte cash for SCSI?

Gaming would be better served by the increased CPU availability than most applications. No amount of super duper mega ATA-5000xxx omgwtf can improve load times like SCSI.

True, but when you add in cost, noise, possibly needing an extra card for the SCSI.


If money isn't an issue use SCSI, if you want to save some cash go SATA II

I go to here usually when i'm looking for computer answers or to ask a question. Not the only computer info forum i've seen, but it is informative imho...Has a motherboard and hard drive sections plus others...

http://www.hardforum.com/ (Had posted link before but was awhile back)

Ok, first, to your question.

The EPoX EP-5ELA3I is a dual AGP/PCI-x board running the 915 chipset, but doesn't support SATA-2. Also, with all the dual boards, AGP support is somewhat limited.

The ASUS P5VD1-X is also SATA-1 but uses the VIA chipset.

The ASRock 775Dual-880Pro is dual SATA-1 and with VIA.

The PT880PRO-A (1.0A) is also dual SATA-1 with the VIA chipset.

I can't find a board that has SATA-2 and dual PCI-x/AGP support. What do you have that generates that sort of bandwidth on a SINGLE ATA channel?

SCSI rocks ATA. A lot.

The way ATA works, your CPU is somewhat heavily involved in the I/O. With SCSI, much less so. So, it's not only faster I/O, but lets the CPU devote more time to other processes.STOP SAYING THAT!

Just STOP!

A decade ago, what you said might have been true. There's roughly 0% overhead with PATA or SATA unless operating PATA in PIO mode. It's all handled off-processor. The only benefit to SCSI now is that (a) there are faster drives available for it, and (b) it has a higher limit to the CHANNEL if you have multiple devices and TOGETHER they are able to break the limit of the interface. If you have a RAID of Fujitsu MAU's, then, thank god for SCSI. If you have a single ultra-fast prosumer drive (like the new 150 gig Raptors!), the SATA interface on any machine is enough.

(Trust me, I have MAUs, Cheetas, and Raptors...)

Everyone keeps demanding these ultra-huge fast pipelines, and they don't have drives that can saturate PATA-100. *sigh*

150 gig raptors... I think I just wet myself.


Anyways, I think the closest anyone came to actually completely saturating the badnwidth for SATA 1 was in a maximum PC article, they tried out a striped array of four raptors, got up to like 130ish range out of 150

But yeah 'max is right the load on the CPU is neglible at best.

A decade ago, what you said might have been true. There's roughly 0% overhead with PATA or SATA unless operating PATA in PIO mode. It's all handled off-processor. The only benefit to SCSI now is that (a) there are faster drives available for it, and (b) it has a higher limit to the CHANNEL if you have multiple devices and TOGETHER they are able to break the limit of the interface. If you have a RAID of Fujitsu MAU's, then, thank god for SCSI. If you have a single ultra-fast prosumer drive (like the new 150 gig Raptors!), the SATA interface on any machine is enough.
You're usually pretty keen on the tech stuff, Max, but this one reminds me of one of your "IE is as secure as FireFox" posts. SATA doesn't come close to matching the efficiency of the SCSI protocol. Not only is the underlying communication less impeded, but even a new Raptor doesn't match a plain old school 15k RPM Cheetah for seek times and sustained transfer rate.

The seek times are important. They're something like 30% faster between a Raptor and a Cheetah (both read and write). Unless you're dealing with single, large, unfragmented files; this is huge. NCQ makes this an even larger advantage for SCSI than the numbers imply.

Just speeding up the clock on ATA doesn't make it automatically make it comparable to a SCSI interface with the same theoretical burst speed. I'm reminded of USB 2.0 vs Firewire. If you buy into the hype on USB 2.0, you'd think it's overtaking Firewire. If you've used them both heavily, you know it's a joke. The SATA vs SCSI comparison is nearly identical.

Open up Process Explorer, copy some files, and watch the hardware interrupts counter. Now do that on a u320 SCSI setup, and not only will the copy run faster but you'll see it won't eat up so much of the Northbridge/CPU. Now try that in a real world scenario, with other applications running, and the difference will be even more pronounced between them.

If SATA was truly able to move data as quickly and efficiently as SCSI, there's absolutely no way that Raptors would still be limited to 10k RPM.

Sadly, for the record, I'm running a 150gb Raptor and a Caviar 400gb data drive myself. I didn't have a slot available for a SCSI card. The Raptor is alright, but my WoW load times are noticeably slower now than they were before on a RAID0 pair of Caviars (Cue someone linking that Anand Tech benchmark "proving" RAID0 ATA isn't any faster than a single drive).

So I had my choice between an adaptec 39320 and a 29160 that were just "lying around" at work;) . I chose the 29160 because the connectors on those cards is unfamiliar to me but the 29160 says it is compatible with 32 bit PCI. I assume it fits in a standard pci slot and the rest of the card hangs over and isn't used? It says it can support 160m/s with the 32bit pci and the 39320 requires the full 64bit pci slot which I don't have and have never seen on a desktop motherboard (only the dell servers we use them in). So, this card has several connectors on it. There is the 80 pin IDE looking one and the trapezoidal shaped 60 pin (I think). Which one is better and what is a good drive to use with it. I am looking for a small size, 18gig but probably 15k rpm. I play battlefield2 and I want to run windows from my ide drive but put the important games on the scsi like bf2 so I can be the first person loaded into the map each round:D .

The 29160 is an Ultra160 based controller. In all likelihood, it'll be enough for anything you'll do with a single drive, but it isn't the most current implementation of Ultra SCSI. The 39320 is an Ultra320 based, and is the latest and greatest.

The trapezoid shaped connector is the "wide" SCSI. That's what any hard drive will use.

The rectangular one is the "narrow" SCSI connector. Slower devices like some tape and optical use that still.

Yes, the "narrow" is physically wider than the "wide".

Sanchek --

Is my MAU and Cheeta at work on u320 faster than my Raptor at home, yes.
Is my pathetic non-SCSI interface killing my processor at home, no.

Of course, if they magicly made a MAU that worked on ATA and a SCSI raptor, the MAU would still be a faster drive, so, what's your point? Of COURSE SCSI is better than (S|P)ATA. But it's not some magic bullet that makes all drives go faster and speeds up your CPU. The CPU isn't involved on ATA DMA transfers except to initate it.

I went looking for drives to go with my card. Fry's has a terrible selection and the only one they have is 36gig, 10k and they want 300$ for it. Pffft. Does anyone have any comments about buying computer hardware, specifically scsi drives from the internet? Do drives come with cables or do I need to buy one? Fry's had some shitty looking cables. Does it make a difference?

Also, is this card going to be somewhat pnp? I just load the small (2mb) driver and it works? Can it run in parallel with an onboard ide drive?

A 147 gigabyte 15k Fujitsu MAU runs you $799 at NewEgg.
A 74 gigabyte 15k Fukitsu MAU runs you $375 at NewEgg.
A 147 gigabyte 15k Seagate Cheeta $1,140 at NewEgg.

10k versions are all about 60% the cost.

10k Raptor 150's are $299.

I bought a drive and now I'm really confused. I got a Seagate st336607lc. This drive has no 5 pin power connector and it has an 80pin trapeziodal connector which I don't have on my 29160 controller. I guess I need to specify 68 pin drive? It is an ultra 320 drive but the salesman told me it was backwards compatible (bs?) There is a 12 pin connector on the front of the drive as well. What is that for? Leds or something? Setting up a scsi drive has really been quite an exercise.

I took back the seagate and got a fujitsu mau3036np. 37 gb, 15k, 68 pin. I hooked it up and it shows as device 15 in the controller bios. I see the controller and the mau in the windows hardware manager. What else do I need to do to get the drive to show up in the windows explorer? Do I need to fdisk and format from the command prompt? I am running the low level format from the controller bios atm.

*Edit: Partition magic saved the day:D

If you're going SCSI, MAU's or Atlases, but I still think you all underestimate the Raptor.

http://www.storagereview.com/php/benchmark/suite_v4.php?typeID=10&testbedID=4&osID=6&raidconfigID=1&numDrives=1&devID_0=279&devID_1=277&devID_2=273&devID_3=309&devCnt=4

Because the Sims2 is where it's at, apparently. Never mind if you actually want to run anything else fast!

Yeah, don't let the fact that it wins 3-out-of-5 single-user performance tests against three drives that should dominate it (but can't) bother you. Also, don't let the fact that in the other two tests, it comes in second --- again, against a field of drives that should dominate it but can't.

It dominates the fastest ATA drives from every company.
http://www.storagereview.com/php/benchmark/suite_v4.php?typeID=10&testbedID=4&osID=6&raidconfigID=1&numDrives=1&devID_0=311&devID_1=297&devID_2=288&devID_3=295&devID_4=301&devID_5=299&devCnt=6

And it still isn't dominated by 15k U320 drives - and you could build a mirror or stripe of them for less than the cost of a single U320 drive.

What test do you use to guage hd performance? With some software I downloaded called HDTach I got these numbers;

Fujitsu MAU3036NP / Adaptec 29160 using 32bit PCI slot
Random access : 5.7ms
CPU utilization : 4%
Average read : 81.3mb/s
Burst speed 117.5mb/s

Those numbers are better than any single drive they have listed in their archive. I know that it smokes my ata133 drive for game play. So it is possible with scsi to buy two identical drives and set them up to interleave between the two and that doubles the data rate? I may have to go buy another one.....

Those numbers look EXACTLY like what they have listed in the archive for a MAU.

http://www.storagereview.com/articles/200411/20041123MAU3147NP_2.html

The MAU (which I have) score slightly slower than the 15k Atlas.

http://www.storagereview.com/php/benchmark/suite_v4.php?typeID=10&testbedID=4&osID=6&raidconfigID=1&numDrives=1&devID_0=277&devID_1=279&devCnt=2

So it is possible with scsi to buy two identical drives and set them up to interleave between the two and that doubles the data rate? I may have to go buy another one.....
That's RAID 0. People will point to synthetic benchmarks that say it's no faster than a single drive for non-server use, but my personal experience has been that it's significantly faster when it comes to game load times.

To do it right though, you'll need a hardware RAID controller. That'll start getting pricey fast for SCSI. You can do it through software (I believe XP has that built in. I know it has a RAID 1 mirroring feature), but that would take away some of the benefit of using SCSI in the first place. It might be worth experimenting with though.

Phelonious, yes, you could stripe them. My point was that you could have striped a couple Raptors for a mere percentage of the price.

Here's the 10 best drives you can buy. (http://www.storagereview.com/php/benchmark/suite_v4.php?typeID=10&testbedID=4&osID=6&raidconfigID=1&numDrives=1&devID_0=277&devID_1=272&devID_2=311&devID_3=279&devID_4=297&devID_5=288&devID_6=295&devID_7=281&devID_8=301&devID_9=299&devCnt=10) The top Ultra-SCSI drives, and the "best" drive from each of the remaining makers (Hitachi, Segate, Maxtor, Western Digital, Samsung and Fujitsu). The Atlas dominates in raw power, and in multi-user performance. Everything else is split between the MAU and the Raptor. The MAU is better and, on an unlimited budget would probably be in all of my machines -- but if price is an issue, you start looking at the Raptor and then looking at the massive price-performance of the Hitachi 7K500 -- which is cheaper than the Raptor and 3x the size.

On a "prosumer" budget, Raptor boot/system and Hitachi data drives can't be beat.

What test do you use to guage hd performance? With some software I downloaded called HDTach I got these numbers;

Fujitsu MAU3036NP / Adaptec 29160 using 32bit PCI slot
Random access : 5.7ms
CPU utilization : 4%
Average read : 81.3mb/s
Burst speed 117.5mb/s
Short test:
Random access: 9.6ms
CPU: 10% (I'm not sure what impact my dual core has on this measurement)
Average Read: 65 mb/s

Long test was the same except CPU was up 2% and average read was down 1 mb/s.

I'm using a 150gb Raptor. Your SCSI drive is clearly better even on an older generation Ultra160 controller. Enjoy those load times.

edit: My big slow drive (one of those newer WD 400gb) only used 1% CPU for the test, in the same system on the same SATA interface. The performance of the Raptor obviously comes with a considerable CPU cost to control it.

I wonder if that's the real issue that made my load times seem slower on the Raptor, moving from my older ATA RAID 0 pair. In a vacuum, the Raptor is faster, but with so much happening that needs CPU attention during a load; the Raptor could be a net loss in those situations.

Your random access times on the 150g Raptor seem...high. My 74g scores 7.9ms, 8% CPU and 62.5 read. Those are within .1ms of the reference time posted in the app - and the app shows 2% CPU as the reference. It might be a failing of BOTH of our drivers -- and not necessarily a problem with the Raptor itself. "Obviously" is a pretty strong word.

I think I should exchange the 29160 for the 39320 at work and then buy a new motherboard...need....more.....power....

http://www.supermicro.com/products/motherboard/DualCore/955/PDSGE.cfm

I think I should exchange the 29160 for the 39320 at work and then buy a new motherboard...need....more.....power....

http://www.supermicro.com/products/...e/955/PDSGE.cfm
The 955x chipset is supposed to have improved I/O handling compared to the 945/915, so that one might be a good upgrade.

The biggest advantage of the 320 is really when you get several devices operating simultaneously on the same SCSI channel, or when you've got a beefy RAID array that can exceed the 160's capacity. So, I don't know exactly what difference you'll see between the 160 and 320 controller. It's my understanding that there are nominal enhancements at the hardware level. That might help you get even better performance in your single drive setting.

Your random access times on the 150g Raptor seem...high. My 74g scores 7.9ms, 8% CPU and 62.5 read. Those are within .1ms of the reference time posted in the app - and the app shows 2% CPU as the reference. It might be a failing of BOTH of our drivers -- and not necessarily a problem with the Raptor itself. "Obviously" is a pretty strong word.
My Raptor's running on the same SATA interface as the 400g drive, using the same drivers. Yet, the 400g drive runs the same test at 1% CPU that the Raptor runs at 10-12%. "Obviously" is a strong word, but it's also appropriate.

I think the fact that there is not a single 15k RPM SATA drive available tells more than any list of contrived benchmarks can. Packaging the drive with an SATA interface would be trivial, and with people paying $600-700 for single video cards on the high end I doubt there would be any shortage of enthusiast demand. The problem is that ATA just plain doesn't cut it at that level.

Again, even HDTach's reference results (select "compare" after testing your drive) show CPU usage at 2% on the Raptor. Since they get 2% CPU on their reference sample, and we get 8% on our samples, just maybe it's not so obvious.

Maybe they ran the reference test on a Cray.

It doesn't say where they got the reference data either. Is that their personal result in the real world, or did they accept a data point from someone else? Personally, I don't put much stock in a test that goes against my real world experience and against the results we obtained using the same benchmark ourselves.

On the HDTach archive list when you click "graph data" and then browse the listings, before you actually plot the graphs at the bottom there is data on the system used for testing. The 74g raptor says it was tested in a xeon system.

It's much more likely a factor of the southbridge or the drivers.
http://www.tbreak.com/reviews/printpage.php?id=291

I see many, many people posting HDTach results showing 2-3% CPU usage, and, again, the reference sample shows 2%. To say the drive is obviously the problem is sheer stupidity.

On the HDTach archive list when you click "graph data" and then browse the listings, before you actually plot the graphs at the bottom there is data on the system used for testing. The 74g raptor says it was tested in a xeon system.
That might explain it. It also says that they were using a Promise controller, rather than the Intel chipset's built in. That probably off loaded more, if it was one of those with its own separate BIOS.

And here's another Raptor getting <1% CPU.
http://www.overclockers.com.au/image.php?pic=articles/170451/raptorsil.jpg

...also on the Intel southbridge. It's the southbridge or the driver man...not the drive.

To say the drive is obviously the problem is sheer stupidity.
Computer sitting idle. Two drives on the same SATA interface, in the same computer, using the same drivers. One takes 10-12% CPU to bench and one takes 1%.

What would you blame it on, if not the drive? There is no other difference between them. Do I need to have my fingers crossed and wait for a full moon to make the Raptor play nice?

Considering everything else is held constant between the tests, it would be "sheer stupidity" to point to anything other than the drive.

Maybe I have a defective drive. Send me yours to test. Oh wait, you get way higher CPU usage in your tests too. Benchmarks in contrived situations designed to optimize their results aren't very interesting to me, when we've both shown that they don't really reflect real world performance.

It's interesting to note that they got much lower performance than they should have on a Raptor. It's slower than my 400g media drive. Are you sure it's a Raptor benchmark? The URL may be a misnomer.

Let me add something here. I am currently playing battlefield2. Before I upgraded the HD I was always complaining about what I thought was latency. In game I would come face to face with an enemy, we would both fire and it would seem like I would get killed with only 2 shots hitting when supposedly you need 4. Sounds like latency right? Well since I have changed the HD I have been kicking ass. I am the one blasting people and barely getting shot once before I get 4 shots out. The only thing that has changed is the HD. I am not sure if the transfer rate has anything to do with it because I assume the entire map is played from RAM. The only thing I can image has changed is the CPU usage. The drive I upgraded from was 12% usage. Regardless of what the statistics show I have done myself a huge favor with this change and it is night and day in terms of game play and load times. If an ultra 160 is good an ultra 320 in raid 0 must be better:devil

BTW: is that supermicro MB the shit or what? I dont see how it gets any better other than maybe being SLI.

*Sigh* You don't get it, do you. I re-ran my test a dozen times, and I score, on average 5% CPU usage. Their reference benchmark shows 2% CPU usage. I show you a link to a guy testing the same Raptor drive on multiple different southbridge controllers and the results for CPU usage change! Let me be clearer. If you take the same drive and use it on different southbridge chipsets, you get wildly different results as to the CPU usage necessary to perform complex tasks - because the southbridge or the drivers that talk to it don't do a good job. Because that's the only variable. The drive was the same in every test.

Same drive, different southbridges, different results.

So, uh, it's the drive, or the southbridge/driver?

http://www.tbreak.com/reviews/printpage.php?id=291

Read that one again. THE SOUTHBRIDGE (and or driver) MADE THE DIFFERENCE. The drive was the same in every test. [And yes, the first tests are single-drive, non-RAID.]

I'm using the same Intel southbridge as those tests you linked. I'm also running the latest drivers from Intel.

I get what you're saying, but it just doesn't make sense. If it were a driver/chipset issue then there's no explanation why my SATA 400g drive on the same channel (even the same physical cable) runs the same benchmarks consistently at one tenth the CPU utilization. The only variable factor is the Raptor.

I did read that the first time you linked it. The whole thing. Including this part:

We formatted the hard drive every time before we switched over to another motherboard to make sure results are consistent.
I'm sure if I completely reformatted the system for each test, I could duplicate those kind of results. As I pointed out before, that really has no bearing on real world usability. Especially when you're talking about something like load times in games, where the CPU is completely maxed.

You never said what's up with that random screenshot you linked. That can't really be a Raptor at <50mb/s.

I dont post very much, but I can say with 100% certainty, that running a single Raptor WD360 is ALOT slower than a RAID0 with the same disks.

I am running Quad WD360's in RAID0 and this is what HDTach had to say.

NVIDIA STRIPE 137.90G
Tested ib 2006-01-31 at 13:25
Random access: 8.1ms
CPU utilization: 4% (+/- 2%)
Average read: 109.3 MB/s

With a burst speed of 228.4 MB/s

This was with the long test.

I dont know if this adds anything, but I read earlier that someone thought that a single drive was as fast as a RAID0 config.

From personal experience, I would have to disagree.

I found it rather interesting that this listed benchmark for Quad WD360's was significantly worse than what I was able to acheive.

HDTach indicated them to be:
Random access; 8.8ms
CPU utilization 29% (+/- 2%)
Average read: 92.5 MB/s

Burst of 110.3 MB/s

here is a ScreenShot (http://www.transportamerica.com/images/hdtach.jpg) of the results if you prefer

That's been my experience with RAID0 too. What kind of controller are you using for the RAID?

Also, when you say RAID0 with four drives, are you talking about a RAID0+1 set, or literally striping across four drives without the parity drive of RAID5 (never heard of that)?

Experience has taught:

Raptors in RAID 0. Yes, its fast. Like anything high proformance, its subject to failure/breakdown. I ditched a 2x Raptor360 RAID 0 due to data corruption and the resulting format. Not much speed was lost by reverting the two drives back to being just 2 indepenant drives. There was a slightly noticable slow down, but now I feel I can sit down and use my computer without worrying if its going to break down a few months from now.

SCSI 320. Its better, but for a home system used for a blend of fun and the kind of professional work you would do in an office environment on a networked PC, you might as well just get some Raptors and give me the difference in price over what you could have spent. SCSI320 is designed for servers. The cards alone cost a fortune and don't fit the typical PC, unless you plan to game on an machine designed as an entry level server. Most of the cards are priced as such. If you want to blow 300 dollars just for a controller card and another grand on hard drives just to make a game a schmidget faster, well. You sir, are an idiot.

It is of course, your money. The US economy likes those who spend lots of money before examining as to why they must. If you just HAVE to have the fastest, well, you can rest assured there is always going to be something mainstream over the horizon that will blow away the technology you figured was "THE cutting edge" at the time and you paid 5x more money for. Otherwise, save your money. You can do just fine on a home system build with a Raptor 74 all by itself (no RAID and no need for even a 150, honestly) for bootup and some slower large drive for bulk storage. Spend the money in other areas, like a better videocard or more RAM or a faster CPU if its just games you are concerned with. Or as I said, just send me the money or flush it down the nearest commode.

Btw, Raptors are still SATA I, because the SATA I and II jump in performance is not that great. The throughput of the drive itself is where the speed lies and not the interface, as far as Raptor go anyway. The drives themselves are the limiting factor. A SCSI hard drive is still just a hard drive, only with alot faster rotational speed than most SATA drives. They cost a fortune because they are designed to run that fiercely for long durations and still be mass produced, and not so much for the interface. If you have ever heard the term "Throwing a hotdog down a hallway" in description of having intercourse with a female that "has been around", you can think of the hallway as the interface and the drive as the hotdog. Anyway, you can get a 400 gig WD SATA I drive with a 16Meg cache for 200 bux at Newegg. A Raptor 150 might set you back 275 dollars. The sexy ones with the see through window will go for 300 and some change. Still cheaper in terms of dollars per gig in relation to speed than the 15k SCSI units.

/shrug. This post is probably so old that you did the build up already and no one really cares.

That's been my experience with RAID0 too. What kind of controller are you using for the RAID?

Also, when you say RAID0 with four drives, are you talking about a RAID0+1 set, or literally striping across four drives without the parity drive of RAID5 (never heard of that)?

I recently bought an Asus A8N-E motherboard for my Opteron 170 Dual Core cpu.

It is based on the Nvidia Nforce 4 chipset and I am using the onboard Nvidia Raid controller.

Unfortunately it does not support RAID 5, only RAID0, RAID1, RAID0+1 and JBOD.

Since I only had 4 SATA connectors, I didnt have the space to run RAID1. I needed more than the 68GB it could provide.

edit : I could buy bigger Raptors, but having just bought a $400 CPU I need to hold off a bit for that.

So yes ... unfortunately I am running in a strict RAID0 accross 4 drives with no protection what so ever.

Ok next round of questions. The adaptec 39320 says it does raid0. When running 2 drives with interleave to double the speed, I assume you have to have 2 identical drives? Do you run them on the same cable or is it better to use one drive per connector on the card?

Dual core CPU's; Does windows run on both cores all the time or does windows run on one core and a game will run on the second core? If you can isolate windows and a game onto separate cores that would be a huge inprovement in performance I would think.

I think you might *have* to use two seperate cords for SCSI, I know you do for PATA. Even if you dont, using seperate cords would be best.

It doesn't quite double the speed, but it will make it faster. However you also run the risk of data loss, so anything uber important I'd have on a seperate drive.

Dula core CPUs, If you use windows 64 bit it takes advantage of it better, but you need applications that take advantage of it, otherwise you wont see too much of a difference, but you can multitask a lot better. IE I encode DVDs while I play WoW with minimal affect on my gameplay.

With applications that do take advantage of both cores you see pretty hefty improvements. It used to take me 30-35 minutes to encode an hour of video to Div-X with a single core AMD 3500+ ,nowwith a 4200+ x2 15-20 minutes.

Ok next round of questions. The adaptec 39320 says it does raid0. When running 2 drives with interleave to double the speed, I assume you have to have 2 identical drives? Do you run them on the same cable or is it better to use one drive per connector on the card?
Yeah, you need drives as similar as possible for RAID0. If they're different, then you'll only get the performance of the slowest one and the capacity of the smallest one.

Definitely use separate cables. Those are two separate SCSI channels, so they can work in parallel better.

Dual core CPU's; Does windows run on both cores all the time or does windows run on one core and a game will run on the second core? If you can isolate windows and a game onto separate cores that would be a huge inprovement in performance I would think.
You can set the affinity of a process to run on a certain core from task manager. If you put your game on the second core and set its priority to high, nothing else on the machine should prevent it from taking up to 100% of that core.

http://accessories.us.dell.com/sna/ProductDetail.aspx?TabPage=overview&sku=341-2478&c=us&l=en&cs=RC959860&page=external

74g 320UW's from Dell for $199 -- Smoking Hot.

The last 320 I got from Dell was a Fujitsu MAU. The other was a Seagate.

Gets better. $175.35

http://accessories.us.dell.com/sna/ProductDetail.aspx?TabPage=overview&sku=341-2478&spagenum=&category_id=5699&brandid=&k=&c=us&l=en&cs=22&mnf=&prst=&prEnd=&mnfsku=&orderby=&searchtype=&pageb4search=&page=productlisting.aspx&instock=&refurbished=

I ordered 2 :)

I'll just re-sell 'em :)

I figured this would be the place to put it instead of starting anew Thread.. Anyone know of a good card for not over $150. One that should be fine playing future games. IE everquest 2

Card? Video card? This thread was about scsi, but I am pretty sure you are not asking about an scsi card for EQ2. If that is the case I think Adaptec has some lower end controllers that might be in that price range, or possibly LSI. If you are talking about video cards I think you are in the 6600 price range.

hehe Ty sorry about that I saw the computer question heading and just assumed someone had a general question MY BAD!

Ok so here is my new computer;

Amd 64x2 3800, 2x512 pc3200 ddr2 cas2
Asus A8N5X
Evga 256mb 7800gtx (470/1200mhz)
Adaptec 29160 / Seagate 373207 LW (65mb/s, 7.9ms, 3%)
SB Audigy platinum

So how do I set this "affinity" of a program to make Battlefield 2 run on the second processor? In general do all of the os/hardware functions like dma and pci run from only the primary processor of a dual system? I am using winXP.

Btw does anyone here play BF2, or are you all still stuck on the rpg's? I am so glad I was able to kick eq. It's the worst drug I was ever addicted to, and thats more than one....

I have a new problem. I was running 2x512mb ddr pc3200 memory sticks. I just bought a second pair of the exact same model for 4x512mb. My 3dmark scores did not change and battlefield2 runs so much better. The problem I have is what apparently happened to my hard drive performance. Here is the before and after picture. It appears that the scsi ultra 160 interface somehow got capped at 38mb/s. Any suggestions?


http://pg.photos.yahoo.com/ph/marcsmith41@sbcglobal.net/detail?.dir=caa2scd&.dnm=7237scd.jpg&.src=ph

http://pg.photos.yahoo.com/ph/marcsmith41@sbcglobal.net/detail?.dir=caa2scd&.dnm=ee90scd.jpg&.src=ph

Only after you added the RAM? Have you tried taking it out and checking to be sure performance goes back up to normal without it?

I took the second pair of sticks out and surprisingly the performance stayed bad. I swear this only started when I put them in. I went through the bios settings and it seemed most of the memory timings changed when I added the sticks. Most notably it dropped the frequency from 400mhz to 333mhz. So I put it back to 400mhz. All of the other timing settings got lower with the 4 stick setup so I left those on the lower setting. I played with the pci bus speed setting, thinking that maybe it dropped down and thats why the bus is limited now. That change didnt do anything. What is the bandwidth cap of a pci slot anyway?

66mhz is the normal PCI speed.

It's not unusual for slower memory sticks to drop the memory bandwidth. It has to operate at the lowest common denominator. It's bizarre that would change your SCSI throughput though.