SSD isn't mind blowingly fast, Actually slower than 3 x Raptor HDD Raid Array

[Berryracer]

I think your problem is very simple to solve and the culprit is one of the following:

1- SSD Misalignment, can be easily fixed within minutes using a small utility

2- Wrongly installed Intel Chipset Drivers without the -overall -overide command so you are stuck with the Windows default drivers for many things

3- no IRST driver

if you want me to hook you up, send me a PM with the TeamViewer ID / Pass.

Everything is monitored in TeamViewer you see everything I do it's not like I am gonna get in there and destroy your pc

 

[GlacierFreeze]

Well, where do I get it? Under what software package is this contained? I checked out Asus support website under P5B Deluxe. I tried the SATA driver and chipset driver but the device manager would not take it.


cheez

Have you tried Intel's site? It's an Intel chipset on the Asus Mobo.

Just go with berryracer's help. I'm sure he'll hook ya up.

 

[Berryracer]

Have you tried Intel's site? It's an Intel chipset on the Asus Mobo.

Just go with berryracer's help. I'm sure he'll hook ya up.

If he has a desktop PC then thats fine but if he has a notebook he must ensure downloading the mobile version of the driver that is compatible with his chipset. just my 2 cents

 

[wotan]

Sort of unrelated, but I'm curious --- how loud are your Raptors? Mine are DEAFENING!

 

[cheez]

I think your problem is very simple to solve and the culprit is one of the following:

1- SSD Misalignment, can be easily fixed within minutes using a small utility

2- Wrongly installed Intel Chipset Drivers without the -overall -overide command so you are stuck with the Windows default drivers for many things

3- no IRST driver

if you want me to hook you up, send me a PM with the TeamViewer ID / Pass.

Everything is monitored in TeamViewer you see everything I do it's not like I am gonna get in there and destroy your pc

Holly crap that's a lot of steps to get SSD working properly. Why did they make things such complicated for SSD???

I did install chipset driver from Asus website for my mobo, but I guess that doesn't mean I have everything installed right..... I will try one more thing to see if I can resolve my SSD performance issue and if that doesn't get resolved I will have you remote in. ^_^

Oh and it's a desktop PC.

Have you tried Intel's site? It's an Intel chipset on the Asus Mobo.

No I didn't.... Good point!

Just go with berryracer's help. I'm sure he'll hook ya up.

I will..

Sort of unrelated, but I'm curious --- how loud are your Raptors? Mine are DEAFENING!

With three of them running together it was loud as MOFO. But only when putting to max load. It was screaming and I actually enjoyed the sound of it. I was getting stimulated. You can tell something is really working. And at the time I had only 512MB system memory (one of the two sticks failed).


cheez

 

[ryderOCZ]

@ Berryracer,
His partition is aligned, it shows in AS-SSD as fine. He is using a SATAII SSD and motherboard, you are using a SATAIII SSD and motherboard, you are going to blow him out of the water on AS-SSD scores.
He is also using a SF based SSD which in AS-SSD scores lower in writes because of incompressible data.
He is also using IAStor drivers, you can see that in AS-SSD screenshot.

@ Cheez,
The AS-SSD specs for the 100GB Vertex2 are here: http://www.ocztechnology.com/images/OCZ_Vertex2_2.5_AS-SSD.pdf
Your reads look spot on, you are about ~20MB slow on writes (Seq). This can be down to driver versions, yes.

Right click "Computer", select properties, then select Device manager in the left pane. Under the section "IDE ATA/ATAPI controllers" take a screenshot or type out what you show/see there.

Thanks.

 

[cheez]

Thanks ryderOCZ. I looked at the bench spec for this SSD in your link. Like you said, mine is almost spot on!

I will report with the screenshot or type them out when I get home! Can't wait!


cheez

 

[Cerb]

Holly crap that's a lot of steps to get SSD working properly. Why did they make things such complicated for SSD???

They didn't. It's really easy in Windows 7, Windows 8, a current Linux, or Windows Server 2008 R2: you click next until it asks for a reboot, then get updates (unless it's a Linux distro which will install the latest packages directly), drivers, etc.. The tuning needs for Windows 7+ are minimal, and often workload-driven (IE, the defaults are ideal for corporate desktop/notebook users, and which other settings are best may vary by what you need it for).

You want to have the best performance on old hardware, with an old SSD, and an old OS, that was made prior to commodity SSD tech being ready for mass consumption. That makes it complicated. It may have support until 2015, but it's still XP.

Put an SSD into a new Pentium system, with AHCI on, and it will not only just work, but the OS installer will handle alignment, defragging will be turned off, Superfetch may or may not be turned off (though it's recommended to do so, in Windows 7, and you'll want to in Windows 8, iff you don't use hibernate), and they made the OS a bit less aggressive with disk access in general (Linux installers also do the right thing, nowadays, and a few popular FSes support TRIM).

I did install chipset driver from Asus website for my mobo, but I guess that doesn't mean I have everything installed

Intel's storage software is not the chipset driver, and is optional (you are using it, though). With a 5-series or newer chipset, you really only need it for features (MSAHCI is almost identical in performance, and is all you need for a single-drive no-RAID no-RST-cache setup). With a supported older chipset with an ICH south bridge, you can often get improved SSD performance by using Intel's drivers (depends on SSD, but there's no harm done if it doesn't).

 

[Makaveli]

well not to make u feel bad but here are the scores of my 2 years old 256GB LiteOn LAT-256M3S SSD

Send me a PM with your TeamViewer ID/Pass so I can take a look at what you are doing wrong your scores are abnormal

Does not compute!

You are comparing his SATA 2 scores to your SATA 3 scores?

Even if his sata 2 drive was performing at maximum it will still be slower!

Secondly for the OP why are your write scores so terrible.

Even my Gen 2 Intel SSD with a 100mb cap on writing does better.

Based on everything i've seen in the last few pages I think you maybe just got a bad drive!

 

[Berryracer]

Does not compute!

You are comparing his SATA 2 scores to your SATA 3 scores?

Even if his sata 2 drive was performing at maximum it will still be slower!

Secondly for the OP why are your write scores so terrible.

Even my Gen 2 Intel SSD with a 100mb cap on writing does better.

Based on everything i've seen in the last few pages I think you maybe just got a bad drive!

oh sorry I totally over looked that. If it's SATA II and it's Sand Force based then there's not much to be done really other than try diff. drivers.

On my previous Kingston HyperX 3K SSD which was sand forced based strange thing I got better results with the Intel Chipset Drivers as the IDE/SATA controller as opposed to the IRST but with every other SSD it was the opposite where IRST gives better performance.

Since is is sand force based I might think that may be the reason

 

[Ketchup]

oh sorry I totally over looked that. If it's SATA II and it's Sand Force based then there's not much to be done really other than try diff. drivers.

On my previous Kingston HyperX 3K SSD which was sand forced based strange thing I got better results with the Intel Chipset Drivers as the IDE/SATA controller as opposed to the IRST but with every other SSD it was the opposite where IRST gives better performance.

Since is is sand force based I might think that may be the reason

not to be nitpicking (sounds like I am though) SandForce is one word, not two.

http://www.lsi.com/products/storagecomponents/Pages/SandForce_Flash_Storage_Processors.aspx

 

[cheez]

Intel's storage software is not the chipset driver, and is optional (you are using it, though). With a 5-series or newer chipset, you really only need it for features (MSAHCI is almost identical in performance, and is all you need for a single-drive no-RAID no-RST-cache setup). With a supported older chipset with an ICH south bridge, you can often get improved SSD performance by using Intel's drivers (depends on SSD, but there's no harm done if it doesn't).

Don't get so nit picky... And you misunderstood what I meant. I told GlacierFreeze that the only place where I got the drivers (Intel chipset for example) are from ASUS WEBSITE. I know the difference between a STORAGE DRIVER and CHIPSET DRIVER... Man you are behind. The direction of the discussion has changed. Read ryderOCZ's post before your post.

not to be nitpicking (sounds like I am though) SandForce is one word, not two.

http://www.lsi.com/products/storagecomponents/Pages/SandForce_Flash_Storage_Processors.aspx

You too.... don't get so nit picky!

LOL



cheez

 

[Berryracer]

not to be nitpicking (sounds like I am though) SandForce is one word, not two.

http://www.lsi.com/products/storagecomponents/Pages/SandForce_Flash_Storage_Processors.aspx

well ok thank u for that clarification. I am a perfectionist so I am glad you corrected t3h m3h :whiste:

 

[cheez]

@ Cheez,
The AS-SSD specs for the 100GB Vertex2 are here: http://www.ocztechnology.com/images/OCZ_Vertex2_2.5_AS-SSD.pdf
Your reads look spot on, you are about ~20MB slow on writes (Seq). This can be down to driver versions, yes.

Right click "Computer", select properties, then select Device manager in the left pane. Under the section "IDE ATA/ATAPI controllers" take a screenshot or type out what you show/see there.

Thanks.

ryderOCZ,

Under IDE ATA/ATAPI controllers in device manager tells the following:

Intel(R) ICH8R/DH/DO SATA AHCI Controller

Is there anything else you need? Should I go to Intel site and look for the latest driver there? Please help. I would like to have my missing ~20 MB/s in write score back.




P.s. I just received my new sound card from Newegg. Holly crap it's fast I ordered it on Monday via 3 Day UPS Select. Got the card installed and is running happy now.


cheez

 

[Zap]

You telling me the OS or the SSD killed my card?? You are just making rough guess, or making stuff up. It would have to be the BIOS that screwed it up.

You missed my point. The card should not have from merely a motherboard BIOS update. Because it did, it should be covered under warranty. Spend $10 or less on shipping, get a working replacement. Easy!

You don't want me to start making stuff up. Could get ugly. D:

Holly crap that's a lot of steps to get SSD working properly. Why did they make things such complicated for SSD???

As Cerb mentioned, they didn't. Using more recent hardware/OS, no problems. just hook it up and install. Heck, it should even work fine without AHCI, just a touch slower.

 

[cheez]

You missed my point. The card should not have from merely a motherboard BIOS update. Because it did, it should be covered under warranty. Spend $10 or less on shipping, get a working replacement. Easy!

I see what you mean now. Well, I can't afford to have long down time. Gotta have working sound ASAP, and I can't stand that crappy SoundMax onboard audio...

Now that I have the working card, I will ... umm send out the bad one for replacement? Oh I will end up having two cards!

You don't want me to start making stuff up. Could get ugly. D:

LOL can't imagine! ^_^



cheez

 

[Ketchup]

Now that I have the working card, I will ... umm send out the bad one for replacement? Oh I will end up having two cards!

Now you can have sound in SLI!

 

[yefi]

....which proves my point exactly. IOPS and throughput are not the same thing.

How can an SSD shine in IOPS but not in throughput when they are related in a fixed way?

 

[AdamK47]

Even new 5900 RPM "green" drives are a bit faster than the original 10 year old Raptors.

 

[Makaveli]

Even new 5900 RPM "green" drives are a bit faster than the original 10 year old Raptors.

I would hope so

The original raptor was using a 1 Single 32GB platter!

The new green drives are using what 500GB platters?

However the access times on that raptor will still be lower.

 

[jaqie]

How can an SSD shine in IOPS but not in throughput when they are related in a fixed way?

Because they aren't related in a fixed way.

 

[Cerb]

Because they aren't related in a fixed way.

You need 10MB, across 50 fragments.

Will having an SSD cause your program to need 50MB for the same work? No. An SSD may provide that data at a rate of 100MB/s, and an HDD at 20MB/s, but unless their performance differences cause the software to change its access habits, the two can have their performance measured by throughput, or IOPS, with a fixed relationship.

throughput = (total_IO_count * avg_IO_size)/total_service_time
or
throughput = (IO1_size + IO2_size + IO3_size + ... IOn_size)/total_service_time

IOPS = (throughput * total_IO_count)/total_service_time (in seconds)
or
IOPS = total_IO_count/total_service_time (n seconds)

Note, as well, that as long as total_service_time is substantially longer than a slower device's seek time, seek times are effectively hidden. The reverse is also true, in that some devices may have fast seek times, but low throughput (such as older CF cards: <20MB/s sequential throughput, but <2ms access for small random reads).

 

[ryderOCZ]

Because they aren't related in a fixed way.

They are absolutely related in a fixed way.

IOPS x (the file size you are measuring the IOPS for) = throughput. Change the IOPS and you change the throughput, change the file size and you change the throughput.

 

[jaqie]

Cerb: As I said several times already, seek times are at the heart of this matter (as is NCQ), as you have just shown and said. Thank you for proving my point for me. SSDs are better at IOPS than hard drives are.

Several experts have also posted on a thread I made in another forum on this issue, and all agree with me as well that IOPS does not equal throughput.
http://forums.2cpu.com/showthread.php?t=99274

 

[jaqie]

They are absolutely related in a fixed way.

IOPS x (the file size you are measuring the IOPS for) = throughput. Change the IOPS and you change the throughput, change the file size and you change the throughput.

If that was true you could come up with a simple formula in which IOPS equals throughput. You can't, unless you are doing direct single sustained linear transfers off of the storage medium. When you do any other kind of data access, seek times become part of the equation as I have said several times before, and as hard drives have a far higher seek time than SSDs, the IOPS and throughput change non linearly. Thus, they are not related in a fixed way, and are not the same thing, as has been posited falsely here several times.



-----------------------------------------------------------------------
to all:
But hey, if you all want to keep proving how stupid you are about this very simple matter, by all means. It's no skin off my back for you to keep spouting ignorant and false facts.

It's not my job to make sure people here understand the truth of things, thank god for that.