500Mb\s?

ah ffs, ppl like u styx, why dont u bring ur excellence and take it up to your standard? ur a waste for actually coming in on a thread with nothing but an ass comment trying to make yourself look good for being thoughtfully better than the rest.

The scenario is simple, I go into the shop looking for a hard drive for my PC. I check out these drives and i read the specifications on the box and it says; 550MB/s Read, 495MB/s Write. Whats it to it? Im good with those numbers, like most people, i take it. I go home, plug the disk in and end up with an ass whooping 10% of the advertised specifications of the drives. 10%. I download a benchmark tool to check what the actual speeds are that im getting from the drives , and yes, 10%, not half, not 60%, not 80%, but 10% of the actual numbers on the box. I think to myself, something got to be wrong.
Nah, nothings wrong, you have case scernarios.

Well, for all of you computer literate geeks and nerds, you probably go and read the specifications and know how to calculate things even before everything is setup, and you get a big grin on your face when you know that people actually think that they are going to get the speed which is written on the box. I know. You already know what those numbers are. Easy Peasy. What an idiot he is going to get 550MB/s reads, lol lol roflmao. For most of us, we don't take the actual type of data, the actual system it has to be running on, if its full or empty, TRIM, and what not into consideration for the drive to be able to output anywhere close to those numbers. Most are probably happy with the drive having other feats like low power consumation, silent running and fast access times.

I know, im not a total idiot when it comes to computers, and i dont go cry about that the drive is actually performing 10% of its advertisement, i dont even expect those numbers written on the box, i live with it, and knowing it will only get slower and worse from the asswhooping 10% im getting now since im missing features on it in the choice of my operating system and other things.

And to sum it up, the main reason that im on about this is not the actually speeds im getting compared to its numbers on the box, its more of a operating system vs another and why the drive is so comparable slower, next to basically the same interface of each other. A driver? A feature? I dont know, its just shitty.

like i was over at ocz, and its like they dont even support the speeds they are advertising their products with. They have probably got tons of weight of people asking why they dont get the advertised speeds and are tired of peoples "stupidity". I showed him a comparison between the disk in XP compared to windows 7, where the same disk are actually performing 3-5x slower because of the operating system, and a return answer that thats just how it is.
 
For good measures i will throw in some benchmarks so people can make conclusions themselves and see what the actual difference is.

Same system, same drives, different operating systems.

Asus P8P67 Deluxe, 3.4Ghz i7-2600k, 8096mb 1600mhz ddr3 cl7.

Vertex 2 50GB, FW 1.0 SATA2 port3
Vertex 3 60GB, FW 2.06 SATA3 port1



Windows 7 64bit, MSAHCI vs IRST 10.5.0.1026;
w764msvirst.png



Windows XP SP2, IRST 10.5.0.1026;
xpirst.png



.
 
Last edited:
The highest i have seen in real world performance is about 78Mb/s, trying to copy a simple file from 1 drive to another and end up with speeds around 60-75Mb/s.... whats up with all these numbers?

one SSD to another SSD? or one HDD to a SSD?

sorry if that was already answered, only read the first post and a few random others. theres a lot of wasted space in these three pages...
 
question about those Win7 vs XP benchmarks, is it the exact same partition? or is a windows 7 formatted partition vs a windows xp formatted partition? windows xp in many cases does not align the sectors correctly for SSDs which could be a possible explaination why you are seeing degraded performance... just a thought.
 
tried several combinations with diskpart and paragon, aligned them at different numbers, the benchmarks doesnt trail that much when u have them aligned properly, anything from 128-4096, doesnt change the benchmarks that much. this last benchmark test is probably the best thought out setup.

fairly happy with the performance of the vertex3 on windows 7, i just dont get it why xp is that slow with them. also the xp benchmarks on the vertex3 was when it was new and probably optimal performance.

one test i did with atto where the disk were aligned by xp is that the benchmarks gets stuck at approx. half the score. alas vertex3 gets stuck 280MB/s when its a 550MB/s, and the same with the vertex2 basically cutting performance in half.
 
Last edited:
dude. i told you before, windows xp doesn't have TRIM support. The moment you run AS SSD more then 2-3 times, your gonna loose a big portion of the performance on the drives. read one of the anandtech articles and you can see the performance differences.

Either way, XP is pretty much out of support now by even microsoft. You have to go towards the future or get left behind with shoddy performance.
 
windows XP is a remnant of more then a decade ago, back when 60gb IDE hard drives were "cutting edge" and we just started to use DDR memory. if you were using windows 98, i would expect you to see similar performance with your SSDs.

couple that with the fact that af22 just brought up, about your sandforce-based drives throttling performance after you write a ton of data within a short period of time, and the performance you are seeing is on par with what should be expected.
 
i TRIM in 7. and the benchmarks should have degraded on the windows 7 benchmarks as well if that was the case after 2-3 runs. its an issue with XP and not the drives. How hard is it to see that for you guys? jesus
 
dek8 - Since the Ferrari analogy (which was so apt) bothered you so much, I'm going to use another car analogy.

When you go buy a car, manufacturers list its horsepower. Plenty of people think "OMG, THIS CAR HAS XXXhp!! It's going to be fast!!!" Then they buy the car.

What they generally don't know is that the horsepower # that manufacturers use is not a practical measure of usable power. It's power at the crank, basically the amount of power the engine can output assuming it's not connected to a transmission or axles, etc.

Now, if you ask any real car enthusiast, they will talk about wheel hp. This is the amount of power that reaches the wheels, and is always significantly less than the crank HP #s.

But that isn't all. Not only is the manufacturer HP rating deceptive because it does not take into account the rest of the vehicle's systems, it also only measures the engine's power output peak. Usable power is what makes a car fast. You don't just want 500whp at 9000rpm, you want a powerband - a significant band of engine speeds where you make lots of power - enough to accelerate smoothly and without spikes or dips in power that can make a car dangerous to drive through corners at the limit.

So, you're saying, how does this relate to my SSD you asshole?
Well, electronics manufacturers have been using the same trick that car manufacturers have been using forever. Just look at a stereo amplifier. They list wattage, but they do all kinds of ridiculous things to come up with an impressive number: "1000watts at 0.1%THD" - 0.1% total harmonic distortion is unlistenable, and not only that, but the only way they can get the THD number that low at the power rating listed is by playing the most efficient tone the can come up with, for a fraction of a second.

Get the trend? Companies list impressive numbers that often have nothing to do with real-world scenarios. They isolate their product and come up with the most impressive measurement possible, regardless of how useless it may be in the real world.

There are so many examples of theoretical #s being used to sell products in the computer electronics market, that I won't even try to list them all, but here's just a sampling.

SATA 3/6
gigabit ethernet
802.11G/N
Ghz (clock speed)
Megabit / Megabyte
Power supply ratings
seriously, this goes on and on

If you're upset because you bought an expensive SSD and it isn't performing at the speed you read on the box, let this be a lesson to you. Most parts don't ever perform at the speeds listed. That's why places like HardOCP can exist and flourish publishing real-world benchmarks. If the advertising on the boxes was accurate, we wouldn't need to benchmark anything, now would we?

The reason I said this thread was a waste of time is because I have a hard time respecting anyone who complains about something which they could easily research and understand themselves, and then ignores the good advice that generous people give take their time to give (I'm referring to the previous posters here, not myself - I am posting out of spite as much as to inform).

I sincerely hope your intent was good, and that you just wanted to further your own understanding. I apologize if I misinterpreted your intent. In any case, you should strive to be more polite when asking others for their opinion, if you want thoughtful, useful answers.

read this!
 
Last edited:
styx, we are past that... its more of a OS comparison... i never expected 500MB/s file transfers on this drive, first of all my other media coming to it doesnt hold that performance so i wont be seeing that anyways... what i meant and my thread starting post was that i was seeing about half an increase from my old HDD's and i was wondering why did they up their specifications by xxx amount of numbers when its really just half a time faster. well u go on and read a little up about it and u end up with certain numbers for certain files in certain scenarios... i know that the advertized speed is more of a design capability and if everything is in optimal conditions and in support it can probably reach those speeds.... now its more of why XP is capping the drives by 3, thats the end result i am wondering about, people come telling me all sorts of things, being this long on a computer im surprised nobody knows the real answer, they come to the conclusion that its TRIM, i was at ocz, they said its normal and they dont stand behind it. i am on xp so thats where i got my initial thoughts about it, i have been reading benchmarks and i knew approx. what to expect before i bought this drive as i was expecting some kinds of numbers when i got my new vertex3 to work beside it, i ended up with certain numbers on xp and i asked myself whats the deal? i posted here and got into that XP is my reason for the low performance. maybe just skip everything i write and just look at the scores, its easier to see the issue im having, and i know you all got an easy fix for it to just throw out xp for windows 7. im looking for a solution to keep my xp and im hoping someone would know what to do with it... ill post a last set of benchmarks for the how effective TRIM is on my drives and maybe people will understand a little better of what the real issue is


The bench to the left is from my vertex2, running on XP and XP only for 8 months, with no TRIM, i have basically used the drive as a standard os disk, with pagefile and all, never giving a thought about its performance and just used it as a regular HDD, i must have run 30 sets of AS SSD, and some 20 runs of ATTO on it before getting the disk into windows 7 to execute TRIM on it.
The bench to the right is the same disk after idling over night in Windows 7 where im sure TRIM has cleaned it up, i go straight into XP run the AS SSD benchmark and grab the photo.


impactoftrim.png
 
Last edited:
here is a relevant tool and thread:
http://www.ocztechnologyforum.com/f...Here-s-a-tool-to-force-TRIM-your-entire-drive

Also, XP vs Windows 7, have you checked the driver on XP vs. 7? Not sure, but it is plausible that some of the performance difference is from a generic HDD driver from 2003 or earlier in xp, vs an updated win7 driver.

Also, for a lot of you talking about XP and not wanting to upgrade; There are lots of ways to run Win7 and have it still look and feel like XP (windows classic theme and disabling the new task bar design) while leveraging the enhanced security and performance of Win7, assuming that your hangup on win7 is the UI. I have done this for several people who love their XP interface at work, and most of them can't tell the difference. (although im sure most of you HardOCPers would easily notice ). I can't see not wanting to use win7 over XP for any real reason other than possibly obscure compatibility issues, however even those can be overcome by using the virtualized Windows XP mode that comes with Win 7. Anyways, to each his own, these are just my thoughts.
 
when you are doing a file transfer inside windows, you will be capped. period. it does not matter how fast your device is.
The copy is read off the source and the written into ram. Then it is written from ram onto the destination disk.
this is the reason for programs such as total commander and whatnot.
the write buffer is the reason for the slow copy speed. the only way around it is to use certain versions of linux (of which i am not familiar, so dont ask) or to use dedicated hardware.
win7 will not write fast enough. it is a limitation of the OS, not the gear.
you guys are looking at the wrong thing. Yes performance is variable with different levels of fill. Yes performance changes once steady state is reached.
NO That is not your problem!
cmon storage gurus, someone should have stepped in with this a LONG time ago. like post number 2.
 
when you are doing a file transfer inside windows, you will be capped. period. it does not matter how fast your device is.
The copy is read off the source and the written into ram. Then it is written from ram onto the destination disk.
this is the reason for programs such as total commander and whatnot.
the write buffer is the reason for the slow copy speed. the only way around it is to use certain versions of linux (of which i am not familiar, so dont ask) or to use dedicated hardware.
win7 will not write fast enough. it is a limitation of the OS, not the gear.
you guys are looking at the wrong thing. Yes performance is variable with different levels of fill. Yes performance changes once steady state is reached.
NO That is not your problem!
cmon storage gurus, someone should have stepped in with this a LONG time ago. like post number 2.


You might want to go back and read my first post in here...
 
Ryan, i do apologize. in at number 3!! you sir are correct LOL

how in the hell did this keep going? ryan told you guys right there. and he provided a much better explanation than I, his is the total answer, perfectly explained.

Kudos ryan :)
 
Last edited:
op- for higher speeds, albeit still limited, do us all a favor and get familiar with Total Commander. it will help somewhat.
 
Not to troll but do you do any actual work/gaming on those SSD's or just speed tests on XP until the drives slow to a crawl with no trim support? Seems like a huge waste of time because those SSD's are working fine, until you benchmark them to death that is...sounds like you just don't like the common sense answers people are posting?
 
here is a relevant tool and thread:
http://www.ocztechnologyforum.com/f...Here-s-a-tool-to-force-TRIM-your-entire-drive

Also, XP vs Windows 7, have you checked the driver on XP vs. 7? Not sure, but it is plausible that some of the performance difference is from a generic HDD driver from 2003 or earlier in xp, vs an updated win7 driver.

How do i check? the drivers butters? Im using IRST on both xp and 7. Tried for hours to find out anything about getting the old microsoft ahci driver for my chipset but im unable to track it down. Could be that the IRST drivers are missing a component on XP or some sort.
 
Last edited:
.... its an issue with XP and not the drives. How hard is it to see that for you guys? jesus
You are correct!

Specifically, the relevant low-level driver in XP. Strangely, there is some "hostility" toward something unique to (probably all) SSDs. This has clearly been addressed in post-XP drivers.

I got curious (once dek8 decided to let us know what OS was involved:)) and did my own tests. I am convinced -- but, unfortunately for dek8, I am not sufficiently motivated to seek a solution. I do wish you luck, though.

-- UhClem
 
cmon clem help me out here, you see that people have no idea here, talking about windows filesystem bottlenecking my 60MB/s when people have been outputting 10 folds of my skimpy 60MB/s for the last 2 decades on their raid setups, yes windows filesystem has to read to ram, what a crock of bs lol, not even a suggestion or thought about what it could be.
 
Last edited:
The key to solving technical problems these days is to realize that you are (almost certainly) not the first person to have the problem.

Google: winxp ssd slow

I have not pursued this, and have no need/plans to do so. But, as I said earlier, I hope you succeed.

--UhClem
 
dek my friend, the answer has been handed to you on a golden platter. Google around a bit my friend and you will see.
there are a number of file transfer programs that fix this. Total Commander, RoboCopy, and TeraCopy, just to mention a few.
Ask yourself: surely there must be a reason for these programs to be widespread.
 
dek my friend, the answer has been handed to you on a golden platter. Google around a bit my friend and you will see.
there are a number of file transfer programs that fix this. ...
No, Computurd, you are wrong. dek8's problem is XP's driver. With a decent (non-constipated) driver, XP has good inter-drive file transfer performance. [The add-ons you mentioned only improve on intra-drive transfers (by using a larger internal buffer, and reducing the amount of head-thrashing).]

--UhClem
 
I have not pursued this, and have no need/plans to do so. But, as I said earlier, I hope you succeed.

This doesnt sound like helping much. You merely have him chasing something he will never acquire.
shall i say..."the world is flat....but i have no need or plans to prove so. good luck finding evidence yourself"
you have 20 posts here, so maybe you arent familiar with the way things work with helping people. that usually actually pertains to offering some sort of tangible help, not making statements and running off.

he might have a problem with XP, of course, thats a no brainer. However, the speeds that he is looking for, he will never get. not even with Win7.


i noticed the OP went back and changed the first post, making it even more confusing for those looking to help. if they read through the post most are trying to explain why his question of;

I even run ATTO and i get numbers in the 400-550Mb/s region, both Read and Write. When do i actually get this speed when working in windows? My system can read and write it but my drive is not able to write at those speeds when im actually copying a file? whats the deal? What about actual file transfers from and to the disk? How can i see this when i work on a system?.

for future reference OP, if your question changes, you should probably either A) start a new thread or B) state somewhere in the body of the thread, not change the first post. anyone who is going to try to understand this now is just gonna be:confused:

i hope that you find your answer OP.
 
i was going to post something similar to clem, im not trying to be rude and im not putting anyone off or ignoring anyone... but computur, simplest way for me to explain this and clem said it, this is a low level issue where things are not talking correctly with each other at the very basic start of the operating system. i have been troubleshooting pc's for i dont remember how long, and its not a fix to launch a 3rd party application on top of a low level malfunction, the applications u mention might help in some areas but it will not really solve the problem im having.

and computurud, im not looking for 500MB/s :), i just want the disk to operate @ windows 7 levels, meaning with a set of controls that uses the disk as its functioning in 7 on this machine. that is 100% possible.

the first post was more of a way for me to whine that windows did not achieve more than x2 the performance of a standard HDD, when in fact i am in xp on a malfunctioning setup. tried to explain that but i dont think people read that lol-

trust me i have used google, what i believe is that nobody with a p67 runs ssd on XP and have actually tested their output, thats why im having such a hard time finding anything relevant to have a discussion with people to sort the problem out, maybe there is a fix, but i cant find it, cant even find a subject to something close to it either.


at least the word is out now, and if somebody runs SSD's on XP on a P67 chipset then they might go test their disk out and report back. possible.


/and yea this thread is a mess lol
 
Last edited:
This doesnt sound like helping much. You merely have him chasing something he will never acquire.
Lesson: The first, and most important, step in "solving" a problem, is precisely identifying just exactly what the problem is.

Once the pertinent info (OS=XP) was revealed, I was able to conduct some tests of my own, and pinpoint "the problem". Now, if I had the XP driver that would solve the problem, I would provide it. I don't know if it exists, but I do believe that OP is helped by my directing him to focus his efforts on that pursuit, and not waste his time with all the other well-meaning, but inapplicable, suggestions being bandied about.

--UhClem
 
i understand.... and what driver do you need? i mean, do you know which driver it is?
My guess is the iastor from intel thats not working good in xp.
Been trying to track down the old MSAHCI driver for this chipset, unsuccesfully, not sure if i can integrate it either.

i got this up on Asus forum as well so eventually someone has to take notice.
 
Last edited:
Now, if I had the XP driver that would solve the problem, I would provide it. I don't know if it exists,

i hope that you find the answer you are looking for OP.
 
Last edited:
i understand.... and what driver do you need? i mean, do you know which driver it is?
No; not only do I not know "which driver", but, as I said, I don't even know that it exists. Consider: Who is both 1)capable of producing, and 2)motivated to do so, an SSD-friendly driver, for an end-of-life'd operating system? (I'm not dumping on you for wanting to use XP; I use XP64 myself on one PC.)

--UhClem
 
So let me get this straight clem.
You went on to test your SSD on XP after u read my benchmarks, right? You run p67? And your findings was equal to mine... If that's so then it comes down to the chipset, possible the nand, if you are using a Sandforce based disk... its either p67+SSD+XP or Sandforce+XP... In either case i don't think we have to ask a single guy to write up a driver/fix for this.
In the user manual of my motherboard XP is still supported and its Asus job to fix that in my case. If its widespread and the word comes out its not that hard for them to fix it, probably a week time if they get enough response on it. If its a sandforce based problem then it can be a little more tricky as i have been in contact with the people at OCZ and they don't give a shit about XP, still, Intel could do something about it.

3 years ago there were 250.000.000 users still on XP, i know it has dropped but there still are a ton, and if this is a P67 chipset issue then im sure it will be fixed, im just surprised no one has noticed.
 
Last edited:
So let me get this straight clem.
You went on to test your SSD on XP after u read my benchmarks, right? You run p67? And your findings was equal to mine...
Not quite. When it appeared that you had demonstrated significant performance differences, for your SSD, where the single variable was WinXP vs Win7, I got curious.

[ Curious -- because I'm an old-time hacker, with a fair bit of kernel and disk driver experience (very early Unix--BillG used my system when he was a student). And I know a "rat" when I smell one :). ]

So, I connected my OCZ Vertex (1) 30GB to a SATA1 port on my NForce4 mobo. In WinXP64sp2 (May '07), I measured (an atrocious) 9.2 M/sec read, and 20.6 M/sec write. A 1TB SATA disk (Samsung HD103SJ) on an equal SATA1 port measured 115-117 M/sec read and write. [These are all sustained transfer rates.]

I then booted an equally-old LiveLinux (Knoppix v5.1.1[Linux kernel 2.6.19]-Jan '07) and performed the same (trust me) tests. Vertex30 got 120 M/sec read and 72 M/sec write. The HD103SJ got 120 M/sec read and write.

Hence, I conclude that there is something "SSD-hostile" in the WinXP SATA driver (ie, the rat I smelled).

I believe that in early 2007, SSDs were not getting any special treatment in kernel drivers (XP nor Linux).

Does anyone running XP get "decent" SSD performance? I would expect so ... but is it with a stock XP driver? Or, are there any driver updates that exterminate this rat?

--UhClem
 
its just reverting back to irst.

and i thought about getting the 60gb version to raid 2 of them up and beat the 120, i wont swap to 7, if i could get msahci running and nothing would change then i will just let it be. guys over at ocz are basically saying the same things, well, 1 guy.
i even believe if i raid them on this setup i wont notice much more improvement, its like something is stuck on this speed.

Good Lord, you mean to say you're still running XP? *facepalm* Why!?
 
This is certainly one of the most amusing threads in this forum. Slapstick comedy, hardforum style.
 
Seriously, as far as I'm concerned this is borderline trolling. He claims to be an experienced user, yet he's still running XP, and surprised he isn't hitting 500MB/sec on an SSD? I mean, REALLY?
 
Either you are...

Wow OP.

Either way, you're trying to use a 10 year old OS with new hardware that it was never designed for. End of retail sales of Windows XP ended in 2008, so manufacturers have no reason to write drivers for it any longer, if you have issues, they're just going to tell you the OS isn't supported and to upgrade. People are giving you the answers, you're just choosing to ignore them.
 
Wow OP.

Either way, you're trying to use a 10 year old OS with new hardware that it was never designed for. End of retail sales of Windows XP ended in 2008, so manufacturers have no reason to write drivers for it any longer, if you have issues, they're just going to tell you the OS isn't supported and to upgrade. People are giving you the answers, you're just choosing to ignore them.

It blows my mind thinking that there are people out there who still think XP is somehow better/faster/"less bloated" than Windows 7.
 
Back
Top