D865PERLL's onboard SATA 1.5 - How to achieve better performance?

[zepherzs]

Originally Posted by ctal

It depends on what you are trying to optimize.

To keep it simple let's say optimize with a balance of all outcomes in mind.
It wasn't a critical question on just one issue, but in general across the board.
However I understand your 10k rpm for access/seek times, and using what HD's we have now for optimizing processes.

Certainly my preference is not to use RAID at all and have it turned Off in my bios. I prefer to use the second onboard port just for occasional but important to me, Disk to Disk transfers when I plug in my duplicate HD, to clone for an identical HD backup, etc...

So back to the meaning of my question in post #14....will my answer be that without using RAID, even if I upgraded to the latest performance MB, and latest performance Processor to go with it, but using my same HD, that my HDTune screenshots results won't be much better?

My original post here as you saw was that I had really thought it was my MB & Processor that was holding my HDTunes readings lower and why I was not able to get even close to the HD's rated 3.0 GB/s and that's why I posted, but you all are telling me that in reality 3.0 GB/s is unreal and the best that any of you get is not more that 1.5 GB/s? ..and so as far as HDTune's readings go with my screenshot, I'm already getting the best that I can and as good as you all have with your newer setups?

Thank you again for your time and replies.

(btw Al... I grew up in Bristol, Conn)

[ctal]

Originally Posted by zepherzs

So back to the meaning of my question in post #14....will my answer be that without using RAID, even if I upgraded to the latest performance MB, and latest performance Processor to go with it, but using my same HD, that my HDTune screenshots results won't be much better?

That's pretty much so.

Upgrading to a mobo with sata2 interfaces would speed up the sustained transfer curve a little bit near the beginning of the drive as I explained earlier.

Upgrading to sata2 would provide virtually no improvement at points further along the curve, where the effects of reduced platter tangential velocity slow the data rate to way below the speed of the sata1 interface. It would also do little if anything for access times, which are limited by mechanical actions occurring within the drive.

Upgrading to sata2 would help the burst rate parameter, for which your measured performance is a substantial fraction of the sata1 interface speed (even more so in the case of my WD640's). But I don't think that would result in much benefit to system performance. I'm not sure but I suspect that the burst rate parameter represents a pure read from cache, and speeding that up modestly I would think would not make a generally noticeable difference in system performance.

Originally Posted by zepherzs

btw Al... I grew up in Bristol, Conn

Small world! I'm in Fairfield County in southwestern Connecticut, where I've been for most of my adult life. Grew up in NYC.

Regards,
-- Al

[zepherzs]

Then what would be the point of me upgrading from my 4 year old MB's which only has an 800 FSB with a 3.4e processor and 2 GB RAM via 1.5 SATA ports to the HD? ..would the point be for just non-HD activity where the benefit is only in real-time processing speed?

To recap and correct me if I'm wrong, but using my above setup with my Seagate Barracuda 7200.11 build of ST3750330AS 3 GB/s rated Hard Drive, where using HDTune gives me 106/86/51 results.. ..keeping that in mind, but as today's date the real-time limitations of the latest 3.0 GB/s for HD technology, that even if I bought the latest Mobo with sata2 interface with the latest 1600 FSB with QuadCore processor technology - that still won't increase my -/86/- GBs average transfer rates to the HD will it? The only thing the new Mobo would improve 'albeit considerably' is any process that does nOt require HD read/write such as real-time processing which is done within the 2 GB of RAM, etc, itself.
Do you understand my conversation flow and do I have the right understanding of this?

[ctal]

Originally Posted by zepherzs

Then what would be the point of me upgrading from my 4 year old MB's which only has an 800 FSB with a 3.4e processor and 2 GB RAM via 1.5 SATA ports to the HD? ..would the point be for just non-HD activity where the benefit is only in real-time processing speed?

To recap and correct me if I'm wrong, but using my above setup with my Seagate Barracuda 7200.11 build of ST3750330AS 3 GB/s rated Hard Drive, where using HDTune gives me 106/86/51 results.. ..keeping that in mind, but as today's date the real-time limitations of the latest 3.0 GB/s for HD technology, that even if I bought the latest Mobo with sata2 interface with the latest 1600 FSB with QuadCore processor technology - that still won't increase my -/86/- GBs average transfer rates to the HD will it? The only thing the new Mobo would improve 'albeit considerably' is any process that does nOt require HD read/write such as real-time processing which is done within the 2 GB of RAM, etc, itself.
Do you understand my conversation flow and do I have the right understanding of this?

I think that where you are going wrong is in overestimating the importance of hard drive transfer rates to overall system performance. On an overall basis, for a given cpu architecture, and considering a typical broad set of applications, the most important parameter is cpu internal clock rate. The next most important parameters are fsb speed and memory clock rate and timings. Those are the major determinants of how fast your system will "feel," and also of how long it will take to perform time consuming processes such as video rendering, for example. If you are playing video games, then the graphics card also becomes a major factor.

You are understanding correctly that the most powerful currently available desktop hardware will not increase your hard drive's 86 mB/sec average transfer rate significantly. It would provide a very minor increase, resulting from the fact that a sata2 interface to the hard drive would be capable of 375 mB/sec (= 3000/8), instead of 187.5 mB/sec (= 1500/8). The fact that your sata1 interface, that is capable of 187.5 mB/sec, is passing data at an average transfer rate of only 86 mB/sec reflects the fact, as I and others have said in this thread, that internal mechanical processes within the hard drive are what limits performance, not the interface. It is NOT correct to think of the hard drive as being "rated" at 1500 mb/sec (or 3000 mb/sec). That rating refers only to the interface. And note that it is 1500 and 3000 megabits/sec, not megabytes. Bytes are what is usually signified by the capital "B" that you used, when you referred to 3GB/s (8 bits = 1 byte, as you probably realize), but the sata2 interface speed is 3 gigabits/sec.

As long as the interface is capable of substantially faster data transfer rates than the "internals" of the hard drive, then it will have only a minor effect on the overall transfer rate. An interface with infinite speed might increase the average transfer rate to perhaps 100 mB/sec for your particular drive, just to take a rough guess for purposes of explanation. The sata2 interface might slow that to 90-something mB/sec, and the sata1 interface to 86. When you have two bottlenecks in series, the narrower bottleneck is the primary determinant of the overall flow rate, and if the other bottleneck is substantially wider, it will also reduce the overall rate, but just very slightly. I hope that makes it clearer.

Regards,
-- Al

[ctal]

Edit to my previous post: To be completely accurate, the data throughput capability of a sata1 interface is 150 MB/sec (not 187.5), and the data throughput capability of a sata2 interface is 300 MB/sec (not 375). That is because of what is known as 8b10b encoding:

8b/10b encoding - Wikipedia, the free encyclopedia

The basic points of my previous post remain as stated, though.

-- Al

[zepherzs]

Hi Al,

Thanks for the reply.

I understand what you are saying about all the issues mentioned.. ..and btw, nice job with the way you explain things!

I know I was sloppy with my bits vrs bytes typos, but we knew what I meant, nevertheless you are right.

I guess really that my results are pretty good considering what I see at the HDTune website on the sorted with highest Speed first - actually they seem low since even mine is good as that.
HD Tune website

(I buy all my stuff at NewEgg too)
I notice your HD link at NewEgg
Newegg.com - Western Digital Caviar SE16 WD6400AAKS 640GB 7200 RPM 16MB Cache SATA 3.0Gb/s Hard Drive - Internal Hard Drives

and the here's my HD
Newegg.com - Seagate Barracuda 7200.11 ST3750330AS 750GB 7200 RPM 32MB Cache SATA 3.0Gb/s Hard Drive - Internal Hard Drives

The reviews on mine were only 4 eggs because of the issues with the previous firmware debacle, but of course mine is the SD 15 firmware which is the 2nd generation beyond that one now so no problem here. If it weren't for that previous firmware issue then it would of been 5 eggs too.

I saw CK8-04's comment about the drives that you agreed upon too
Question about WD Raptor Replacements

I'm not really sure how many platters I have or not on my HD
Seagate Technology - Barracuda 7200.11 SATA 3Gb/s 750-GB Hard Drive

Anyway, I have what I have.... and I have two of them!

At some point I will update my MB and processor, but for now I am here.

I guess this is as good as I can get with what I have...

Well thanks Al for all your replies, I have enjoyed our conversations.

Sincerely,

[ctal]

You're welcome! Looks like you've done some pretty thorough research.

Yes, you have an excellent drive. I suspect it is three 250gB platters. The drive benchmarks at the HDTune site look slightly outdated; I didn't see anything there that was larger than 500gB. All else being equal, a higher capacity drive will have a faster sustained transfer rate than a lower capacity drive, because it has a greater data density (more bits per square inch), which increases the number of bits passing under the heads per unit time.

Regards,
-- Al