DMA66 cable doesn't seem to improve performance at all

[Win_98]

vs using my old regular 40 pin cable. Hmm can someone explain.

[Target]

Just checking, so don't take offense, but the hard-disk is UDMA 66 right and is enabled in the bios as such?

[canit]

What kind of benchmark tests did you run with the old vs. new cable? I would think there would be more throughput with the 80pin cable in a burst test. I don't think you can have UDMA66 with a 40pin cable. Either way there shouldn't be that much difference since most HDD's can't sustain more than 33MB/sec throughput.

[Roy]

ATA66 cannot work without the 80-wire interconnect. There is an extra "ground" wire between the signal wires for shielding of the higher frequency pulses. An ATA66 cable, by itself, will not change the performance in a 33MB/s system.

[Graham]

Someone posted in another topic that you if you had two devices on the same ide ch, one of which was 33 and the other 66, then they would both run at 33, does anyone know if this is true ??

G

[seti]

like Roy said...

An ATA66 cable, by itself, will not change the performance in a 33MB/s system.

Too being actully using the ATA66 standard you need an ATA66 compatible chipset, or controller card, and an ATA66 Harddrive. Even if you had these things, you might not see any improvement, because the ATA33 standard is just becomeing a bottleneck for the fastest ide drives.

[Roy]

Graham ~ as I understand it, that is not the case. Promise says their card supports 4 IDE devices and implies mixing of standards is OK.

[Win_98]

I have DMA66 controller, when I boot up with 40pin cable it say 80 wire not connected and still show as dma/66 upon bootup. I benchmarked in both 40pin and 80pin cable, exact same throughput! and exact same benchmark! you tell me, they are pulling our leg with DMA/66 ****? About 10,000 score in sandra harddrive bench and about 13meg/sec throughtput read, 7 meg/sec throughtput write. It look like this 30gig 5400 rpm drive I am using isn't even using full dma/33 potential! so I guess the DMA/66 thing is just to make this drive sound cool and up to date?!
[This message has been edited by Win_98 (edited 09-29-2000).] You know the funny thing is that they even make 8gig with dma/66 capable but does not improve performance at all vs dma/33 since the drive itself only uses DMA/16 full potential at best hence another coverup by the manufacture to make you think it is faster.

[This message has been edited by Win_98 (edited 09-29-2000).]

[NDC]

DMA66 and DMA33 doesn't make any noticeable differences in disk performance. That just refers to burst speeds. A for 5,400rpm Vs. 7,200, 10,000, or 15,000rpm, you will see eye-popping differences! http://sysopt.earthweb.com/forum/smile.gif

[techs]

ata66 was designed as a future standard, and no drive will show appreciable difference in performace with just ata66 enabled. the fastest ide drive cannot get above a sustained 25mb rate even in an optimized test(adjacent data,etc). once the marketing boys(and girls)saw it was selling, they decided ata100 would sell even better. the previous post was correct, if you want improved, noticeable hard drive performance, get a faster drive in rotational speed.
As to the mixing of ata-33 and 66 the latest i have heard is if you have the master as ata66(with the 80 pin cable, etc) that drive is capable of ata66, while the slave drive will be ata33(if thats what it is, or 66 if thats what it is) Before every one jumps on my case, ata66 has a very small benefit in burst mode, but since the disk cache is at most 2mb figure it out for yourself-2 mb at 33mb per second versus 2mb at 66 mb per second. millisecond increase at the very most.

[Fingers]

That is consistent with what I have read also, techs.

An article at Seagate states "Yes,...a legacy ATA specification drive can coexist with an Ultra ATA/66 drive, and both devices can attain their maximum performance."

The Quantum white paper on ATA/66 implies that using an ATA/66 cable can improve data integrity even on ATA/33 systems.

[NDC]

One of my system hasa Soltek EV-1 M/B and that hasn't been the case. When I tried attaching an DMA33 drive on the same channel as the DMA66 (Conductor No. 80 cable), it just dropped the DMA on the DMA66 drive to DMA33 on the boot screen where the DMA is shown. Am I missing something in the setup?

[techs]

I have heard of that NDC. that is the reason i qualified my remark. i wonder if that is unique to certain boards, or has something to do with any cable select devices, or perhaps even position of ide device on the cable?

[Fingers]

I also have some doubts about being able to mix ATA standards and retain maximum performance on both.

ATA/66 mobos have a special circuit on line 34 of the IDE cable. If the line 34 circuit is open, then the bios will assume that an ATA/66 cable is being used and attempts to configure the IDE channel for the higher standard. But the bios also has to detect the presence of an ATA/66 drive, and I am not sure how it does that. IF it is by leaving the line 34 circuit open, then wouldn't installing an ATA/33 drive on the same channel just close it, thus ATA/33?
Maybe that's not how the bios determines the ATA standard for the drive.

NDC check out these links where I have gotten some of my information/mis-information, maybe we can get to the bottom of this. http://sysopt.earthweb.com/forum/smile.gif
http://www.quantum.com/src/whitepapers/wp_ultraata66.htm
http://www.seagate.com/support/kb/disc/ultra_ata_66.html
http://www.wdc.com/products/drives/drivers-ed/ata66tp.html

Also of possible interest;
Why is there a blue, a black, and a gray connector on my Ultra ATA/66 cable?

The 40-pin, 80-conductor cables support the Cable Detect feature. The blue (end) connector should be connected to the system board. The black (end) connector is known as the master position. The gray (middle) connector can be used for slave devices. Although an Ultra ATA/66 device should work fine in either the black or the gray connector, if you are having difficulty, consider troubleshooting with the drive in the black connector.

[Ygor]

Maybe this is why the DMA gets dropped on my WD 7200 rpm ata66 while my older Maxtor (ata33) DMA works fine and beats it?
This happens on a bx6-2 and a bx master board.

[Win_98]

The good thing about this new maxtor VL series huge size drive is that they are noiseless. I couldn't even hear a churp when the drive is seeking. Unlike my last 17.2 gig I got a couple of month ago, produces noise and are only half as fast scoring 5000 on sandra hard bench. I suppose the drive capable of DMA 100 but only show up as DMA/66 because that the highest my controller support. I am very please with it so far, in the future I may buy more of this kind hopefully it will last alot longer then their previous model.

[Sweeper]

My Maxtor Ultra DMA66 7200 RPM get's blown away compaired to the Ultra66 25 gig in Sandra. According to Sandra my HD (which is hooked up to the Ultra66 controller) performs just as good as a Ultra33. ?????????????????????????????????????? Using an Abit BE-6 MB

[NDC]

As I have posted numerous times in this forum, connecting a DMA33 device on the same cable as the DMA66 will bring down the DMA down to 33. I did read the Seagate article about DMA66 drive attaining maximum speed when attaching DMA33 on the same IDE cable with a Conductor No.80 IDE cable. But that just doesn't seem to make any sense, of course this not making any sense is in my opinion. But there is a fact that I learned from experimenting with numberous systems, that is DMA33 + DMA66 = DMA33. To me that seems to make perfect sense though, because System BUS moves at the speed of the slower BUS device AFAIK.

Ex:

PC-66 + PC-100 = 66Mhz
PC-100 + PC-133 = 100Mhz
DMA33 + DMA66 = DMA33


Well, anyways. The important thing is that, I don't mix my DMA33 drives with my DMA66 drives and that has kept me clear from that problem. http://sysopt.earthweb.com/forum/smile.gif




[This message has been edited by NDC (edited 10-01-2000).]

[Ygor]

Sweeper, is that with DMA enabled? I now have mine on separate IDE's, no change in DMA status or performance.

[otheos]

Sorry NDC, but as long as you use an 80wire cable and the devices are at least EIDE each drive will work at its speed. i.e. the ATA100 will work at 100MB/s and the UDMA33 at 33MB/s. This is normal, because when the UDMA33 device stops using the bus, the bus resets to the drive that is using it.
I have personal proof for this, more to what the big HD companies already say (and at least VIA -don't know about intel/ali).

However and old ATAPI (CD-Rom/Zip) drive that use PIO mode are taking over the bus and the bus cannot revert to UDMA after their use. So you need minimum EIDE (also known as UDMA1 i.e. 16.6MB/s in DMA mode not PIO4).

EDIT: NDC, although the memory bus indeed reverts to the lowest speed memory and stays there, this is not the case for the IDE bus for a simple reason: it is not accessed by both drives simultaneously at any time. So whichever device asks for the use of the bus, the bus reverts to the device's speed. That is not the case with memory where all sticks occupy and use the same bus simultaneously.

[This message has been edited by otheos (edited 10-01-2000).]

[NDC]

No, need to be sorry, Otheos. http://sysopt.earthweb.com/forum/smile.gif
It just really sucks that what you're saying and what Seagate site posted hasn't been the case for me. I kinda get the feeling that certain M/B just don't seem to work that way, after all M/B's are at least a year old. What do you think? Do you think also that it has to do with certain M/B's? I've tried eevrything there was in the book to attain the maximum DMA speed on both IDE HDD's. Whenever I connect the DMA33 on the DMA66 drive, DMA66 just drops to 33. http://sysopt.earthweb.com/forum/frown.gif Any good suggestions, Otheos?

[otheos]

I noticed that you said that when you connect the UDMA33 drive on the 80wire cable along with ATA66 drive, the BIOS (on the POST screen) sees them both as UDMA33.
This is where the problem starts. I was about to tell you (to support my case) that if you see the bios when it POSTS it sees each drive with its speed eventhough they are on the same cable and have different speeds, but it seems that in your case this doesn't happen.
What chipset have you got? it's a dual mobo you got right? and if you talk ATA66 it must be the MSI right? so I guest it's the VIA133 chipset huh? But it makes no sense. Even my old MVP3 (SS7 mobo) chipset could tell the two drives had different speeds eventhough they were on the same cable.

[NDC]

I appreciate you to take the time to reply back to my question. http://sysopt.earthweb.com/forum/smile.gif

Actually this for my other system, My son's. It's a Soltek SL-67EV1, Apollo Pro 133 chipset.

Here is the link to the mainboard that I'm refering to:
http://www.soltek.com.tw/English/eng_main.htm

[voogru]

Ya Know its funny that Best Buy Is selling 80 Pin Cables for $30.00 and im getting 40 pin cables for $2.00 and on the box of the 80 Cable it says "It will work for DMA/33 and DMA/66 and it will Increase Data Transfer" but it seems that its not showing improvement. So what would you rather do?

Spend $30.00 ON a DMA/66 Cable that wont Improve perfomance.
Or
Spend $2.00 on a cable that wont improve performance either but its cheaper.

voogru

[This message has been edited by Bill Gates(edited 2-16-1981).]

[otheos]

$30.00!!!!?

they are £6 at maplins over £3 for the 40wire ones. I would presume they'll be normally priced anywhere else.

[NDC]

I'm already using the Conductor No. 80 cable. Any idea why I'm not attaining the maximum speed on both DMA33 & DMA66????

[Fingers]

The information below doesn't directly address your situation NDC, but it might provide a starting point to work from.

From the Quantum White Paper.(see link above) Use of an 80-conductor cable also has enabled Quantum to double the Ultra ATA burst data transfer rate -- simply by streamlining data management within the drive. Specifically, the new cable specification makes it possible to halve setup times prior to signal read operations, which occur at the beginning of each burst data transfer.

Before the drive controller reads a signal, it must "look" at the data lines to determine whether they have switched to a high or low state (i.e., a one or a zero). If the controller examines the lines too soon after the previous data burst, they might still be somewhere between high or low. Hence the controller must wait for the lines to "settle down" to ascertain the current state. That waiting period is the setup time.

The addition of 40 extra ground lines to the Ultra ATA cable spec considerably reduces signal crosstalk and ringing between the data lines (Figs 2 & 3). That allows the lines to "settle down" much faster, thereby slashing setup times in half. And that is what enables the Ultra ATA/66 interface to transfer data at twice the Ultra ATA/33 rate without requiring any other significant changes to the Ultra ATA specification, especially to the DMA protocol

What are some common troubleshooting steps?

...Some system boards may not successfully handle Ultra ATA/66 on both ATA channels. If you have difficulty, consider troubleshooting with the device in the Primary Master position.

I don't know, but I figure the more information I throw at the problem, the more likely something useful can be found http://sysopt.earthweb.com/forum/wink.gif

Just so we can rule this out; Is the ATA/66 drive on the black connector and jumpered as the master drive, and the ATA/33 drive on the gray connector and jumpered as slave? Are they on the primary channel?




[This message has been edited by Fingers (edited 10-02-2000).]

[Win_98]

there is no need to buy ata66 cable, most come with a motherboard that support dma66 and harddrive as well. for 30 buck, best buy can shove that back down their a** that too much to pay, and hell I think it only cost them 4-5 bucks?! making 25 buck off you. I think best buy suck at selling computer component, rip off places if you ask me. you know the funny thing is most of us buy from them but you got to be careful on certain thing they sell, because some of it is way overpriced. I wouldn't buy sound card from them, wouldn't buy video card, wouldn't buy motherboard, wouldn't buy **** from them anymore because they are retailer, not discount computer store. Most harddrive with DMA/66 would peak out at 21-25mb/sec so dma/33 is more then enought for it. As for Faster drive such as 7200 rpm and really huge 60gig or more, have it begun to make some use of dma/66 capability?