|
thank you sir
|
|
|
|
|
|
thank you sir
|
|
|
|
|
Was hoping the hardware gurus could provide some insight into this comparison.
dual 15,000rpm drives setup with raid-0
vs
single SATA 10,000rpm drive with databurst cache
These will be used on a development machine for code compilation.
|
|
|
|
|
I doubt that RAID-0 of two 15000 rpm drive could be beat by a single drive.
Farhan Noor Qureshi
|
|
|
|
|
Well, I figured the dual raid-0 drives would be better. What I'm wondering, though, is how much better. i.e., if its only a minor improvement, it might not be worth the extra money to get the dual raid-0 drives.
|
|
|
|
|
|
Dave Calkins wrote: dual 15,000rpm drives setup with raid-0
vs
single SATA 10,000rpm drive with databurst cache
Depending on the drive... but generally single SCSI U320 single 15k runs at 87 MB/s, a RAID0 pulls about 140+MB/s. My fastest RAID0 was a 4 disk RAID0 of 15K U320 drives. It's average speed was 280 MB/s, and was a hardware RAID0.
A single 10k SATA runs at about 78 MB/s with RAID0 starting at 120 MB/s and going up.
_________________________
Asu no koto o ieba, tenjo de nezumi ga warau.
Talk about things of tomorrow and the mice in the ceiling laugh. (Japanese Proverb)
|
|
|
|
|
Just purchased:
* AMD Athlon X2 64 3800
* 450 watt power supply
* Corsair XMS2 TWIN2x1024A-6400 (2x512)
I put these components together using the stock AMD heatsink and fan. These are the only components attached at this point other than a keyboard and speaker and power button.
When I try to boot this to get to the BOIS, it just makes repeated long beeps (or could be a long beep and short beep very close together as a get a click in the middle of the long beep). The manual says it is a problem with RAM or MB, but they are new and have been reported as compatible. I tried a stick of KByte 512 DDR2 PC2/533/400 and it did the same thing.
The power supply was first connected by the 20 pin connector and the 4 plug 12v connector for the CPU. After this did not work, I tried adding the four plug extension to make the 24 pin plug and still no change.
Any ideas of what it could be besides dead parts? It is not like there is a lot of components, just MB, RAM and power.
Is there any way to determine the CPU is install correctly and functioning? The CPU dropped right in and seem to go easy.
|
|
|
|
|
Rocky Moore wrote: Any ideas of what it could be besides dead parts? It is not like there is a lot of components, just MB, RAM and power.
Either youi've got the RAM in the wrong slots, for instance in slots 3&4 instead of 1&2, or the RAM and/or sockets are bad. If replacing the memory does the same thing, you've probably got a bad motherboard.
I had two Gigabyte motherboards in last machine I built. Both of them exhibited problems where the machine would just hang a random. After replacing everything else in the machine (CPU, RAM, Video), I put the original stuff back in the machine and swapped the motherboard with an Asus. Worked like a charm. The machine has been running for two years straight without a single crash or lock up.
Dave Kreskowiak
Microsoft MVP - Visual Basic
|
|
|
|
|
I dread the thought of it, but I am thinking it may be the motherboard. Just wish there was a way to know the CPU is working correctly as it has a short window of RMA.
I first looked at the ASUS MBs but they did not have the features I need such as 4 IDE devices (have 3 I need to use) and 6100 built in video. Was amazed to see some MBs did not even come with LPT ports since one of my printers still only works via LPT.
Over the last three weeks I have dug through tons of information trying to what works with what, what is just overpriced and what is garbage. Additionally, what will work with Vista or at least to a point.
This upgrade adventure has at least taught me that the designers out there for hardware need to be whipped daily as the industry has become far to complicated. In the old days, you needed RAM, you purchased some of the speed you needed and presto it worked. Now it depends on vendor, volts and other things. Then there is the different CPU socket types and chip sets, along with different cores. In the mix we have to included the diffent models, where there are several for each brand and before you know it you a MESS!
I am thinking over two decades ago, I should have dug a hole and through my computers in it and just became a farmer
|
|
|
|
|
Rocky Moore wrote: Over the last three weeks I have dug through tons of information trying to what works with what, what is just overpriced and what is garbage. Additionally, what will work with Vista or at least to a point.
I'm waiting for Quad Core processors to come out before I build my next machine!
Dave Kreskowiak
Microsoft MVP - Visual Basic
|
|
|
|
|
Dave Kreskowiak wrote: I'm waiting for Quad Core processors to come out before I build my next machine!
Same here. I'm also planning to wait until mid nextyear to see if the architecture changes AMD is releasing with it's quadcore design will surpass intel again.
|
|
|
|
|
Yeah, I am figuring once Vista settles after the first of the year, I will probably get a new high performance machine for development and give this to the wife. Hope I getting this one running by then
|
|
|
|
|
Rocky Moore wrote: Was amazed to see some MBs did not even come with LPT ports since one of my printers still only works via LPT.
PS2, COM, LPT, and Game ports have been legacy interfaces with no new hardware for years. The shedding of the ports has been starting on desktop boards for at least a year. Laptops have been shedding PS2 ports for years. Paradoxally those are the ports with the greatest desktop longevity.
Rocky Moore wrote: In the old days, you needed RAM, you purchased some of the speed you needed and presto it worked. Now it depends on vendor, volts and other things.
Unless you're having wierd compatability issues (or overclocking) none of these should matter. All DDR1 runs at the same voltage, all DD2 at the same voltage. Matched sets are basically marketing BS other than the fact that modern cpus use twin sticks to double bandwidth, which has happened in the past as well.
Rocky Moore wrote: Then there is the different CPU socket types and chip sets, along with different cores. In the mix we have to included the diffent models, where there are several for each brand and before you know it you a MESS!
You had diffent cpu types and sockets every few years in the past as well, along with mobos that had the same socket but didn't support chips beyond a different rev. Aside from the fact that AMD started designing it's own sockets (initially to avoid playing catchup with intel) nothing's changed.
|
|
|
|
|
dan neely wrote: all DD2 at the same voltage.
Actually, no! They range from 1.8v to 2.6v and on some systems is quite sensitive (at least at the 800 PC26400 level). There have been numerious problems with different motherboards on Newegg regarding this issue.
|
|
|
|
|
The specification is 1.8V. Higher voltages are used by idiots trying to overcome the fact that at higher-than-rated speeds (overclocked modules), there isn't enough time for the outputs to stabilise at the level required to detect a value transition. I can't honestly remember whether a high voltage represents a 1 or 0 bit these days.
Using higher-than-rated voltages means that the inherent R/C (resistance/capacitance) junctions within the circuit will take less time to charge and will therefore stabilize more quickly. It also means that more power will be dissipated and therefore more likely that the chips will overheat and melt the very thin metal traces joining parts of the circuit, meaning failed chips.
|
|
|
|
|
Hi all,
I am really frustrated with the SCSI drive r/w speeds on our systems. One system is having Adaptec on-board controller with a Fujitsu 15k drive connected to it and the other one is having LSI logic on-board controller with a Seagate 15k drive connected to it. They both are giving me r/w speeds of 14MBps to 17MBps which is not acceptable with a SCSI drive. I have tried a Maxtor drive in place of Seagate and that too is having the same speeds. But strangely the Fujitsu connected to adaptec controller is acting differently, I have downlaoded a scsi diagnostic utility from Fujitsu website and whenever I run this utility, the r/w performance boosts to 70MBps. But this utility can not be used on other drives (only on fujitsu drives).
You can download the speed utility program that I use to check the drive performance from here
HARDDISK.exe
Can someone help me out with whats going wrong over here?? I even updated my aspi drivers to the latest version (V4.71). What speeds are you getting on you SCSI drives in non=raid option??
thanks,
-Pavan.
|
|
|
|
|
All of my servers are in the the range > 14MBps to 34MBps (more or less with your software.),
All of them are either LSI or Adaptec SCSI. But all are Maxtor Drives.
(I never had any good luck with Seagate.)
I don't know where your getting the 70MBps, but I didn't.
Sorry, Not much help..
Progload
|
|
|
|
|
I'm getting 70MBps using a Fujitsu drive and only after I run the SCSI diagnostic tool provided by Fujitsu. But it works only with Fujitsu drives though... at present it is on-board adaptec controller, I haven't tested it on other cohntrollers.
-Pavan.
|
|
|
|
|
Could be the Fujitsu software tools "communicate" right to the drive controller on the drive and do somthing, and not to the SCSI controller.. humm.. don't know.. but it's interesting.
I did notice your software produces a different speed average each time it runs but it was mostly in the ranges I gave you.
Wish I had a Fujitsu drive and time to dink around with it..
Don't think i'll be much help, Sorry.
Curiosity.. where did you get the diag tool?
Progload
|
|
|
|
|
I got it off Fujitsu Europe website at that time... here it is http://www.fujitsu.com/us/services/computing/storage/hdd/support/utilities.html the SCSI utility - now its available on US website too..
I think SATA is far better than SCSI in my case. I built a new system with SATA drives in RAID0 and I am getting around 121MBps (using WD Raptor 10K SATA drives).
-Pavan.
|
|
|
|
|
Yes, SATA as in most cases would be a better option I think.
I only do Scsi, if I really need or want it "Bullet-Proof-No-Fail"
Sorry I was asking about the Spectrum Speed tool..you had a link too..
I couldn't find it on google..
Progload.
|
|
|
|
|
That tool is specially created by the Spectrum GmBH corp (strategic-test) to test their systems while testing their high speed PCI and PCI-X Signal acquisition and AWG cards. So, its not a commercially available tool.
-Pavan.
|
|
|
|
|
I see, ok, well thanks.. It's an interesting little program anyway.
If I come across a fujitsu scsi drive, in the near future I'll drop you an email on what I find out.
Wish I could be of more help. It's interesting issue..
Progload
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
