SSD Thread (Benchmarks, Brands, News, and Advice)

[max420]

he should notice some improvment with the SSD, even though its a lower clock, it's still a core2duo, and even if it is just with boot times and app launches, it will be more responsive.

my friend has my old D420, CoreDuo, that he put a 64GB SSD into, and while it encodes video at the same speed it used to, it index's, and muxes files wayyy faster than when i had a mechanical HDD in it.

[ronan_zj]

i am shooting 1Dmkiii and 5Dmkii, and using Adobe Lightroom and Canon DPP. if you are using SSD, the bottle necks are ur CPU and memory.
IMO, at least you need 8G memory. If you are buying a SSD, you should consider about Intel SSD or OCZ vertex/Turbo, it gives u good 4k and 512k random read/write. As I have already stated, u also need more memory and fast CPU.

[vostro1400user]

You know, this is why benchmarks are so questionable to me - when they're believed at face value without any real, real world experience to back them up.

First, for the 17.9 Win 7 boot up times...

I can't read/understand German, but I can see that they tested the drive under Vista x64 - so then they Award it because of its Win 7 boot time? Something does not sound right there? (I'm probably way off, but that is what I gleamed from trying to see the testing methodology they used).

As for the bit-tech.net article quoted above, hmmm....

I understand that they are testing a G1 Intel (not defending Intel here, just that if Intel was a little better in Maximum combined read/write response times - because it thoroughly trounced everything else in Average response time and combined read/write Speed (throughput - MB/s) tests, it would be the 'best' in every category of combined read/writes). I wonder what a G2 with the new firmware would do in these tests? As good as the G1 is, the G2 certainly is better, I believe.

More to the point, I wonder what a 160GB G2 would do. That would be a much better comparison to the 256GB Corsair, right? (The bigger the SSD drive, the faster it (usually) is...).

Also, though they claim to do real world testing - they don't. I don't clone my installations because they are not as good as a clean install - they cloned, which makes the SSD's especially look better than they are, for a user like me. Then, they tweak and disable almost everything in Vista and use that as a real world boot up test? Not one of the more useful sites in my book, sorry.

I want real, real world tests; that would mirror my experience with the products if/when I decided to buy them and use them in a real production situations - not an artificially created environment that showcases a single good or even great point about an obscure and possibly insignificant benchmark metric. And, a metric that they failed to show how it would even affect a user in a real world setting.

Don't get me wrong - I'm not disputing the actual numbers. What I am disputing is the fact that they are pulling at straws to find something good to say about a product they're reviewing, but failing to convey how its supposed strength will benefit the user. In that respect, this review is a fail - not only now, because the G2 has been released since, but also at the time because they failed to test it against a 160GB G1 and also show how the benchmarks they produced related to the real world.

I also had to chuckle that a full format kills the performance of the drive - all that is (a full format), is simply writing to every cell - so, so much for no significant performance degradation, huh? But they masked that fact well, with benchmarks as 'proof'.

win 7 bootup time is from ocz site rather than that review. for real world test, it's a pity we haven't seen real scientific test reflecting all aspects of real world performance, but only some partial test for specific aspect. since stuttering will be big issue for my applications, i just more concern on this. according to that review, Random Combined Read/Write Response Time (Maximum) seems to be more critical on this issue, although i don't know whether this is true or not. as that review shows this random combined response time on intel G1 is nine times longer than P256, and five times than vertex, let's assume G2 gets significant improvement, (write speed doubled than G1 from other review) can we assume G2 improve about 200%(i.e. this number would be half of G1), then it's still way more than double of vertex number and four times of P256 number. but we never know the truth on this regard for G2 from available online resources so far...

for that full format issue, as long as GC or TRIM works, it's not a big issue i guess.

[hankaaron57]

hankaaron57,

sure they'll degrade, but by six times?

I'm not suggesting to do a full format, just commenting that a full format is effectively writing to every available cell and the Samsung's/Corsairs are supposed to offer consistent performance from when you start using them, till, I guess forever - because they're supposedly shipped in a 'degraded' state to begin with.

The full format 'bug' sure throws a wrench on that view, huh?

Roger that, just verifying.

[tilleroftheearth]

Roger that, just verifying.

Cool!

win 7 bootup time is from ocz site rather than that review. for real world test, it's a pity we haven't seen real scientific test reflecting all aspects of real world performance, but only some partial test for specific aspect. since stuttering will be big issue for my applications, i just more concern on this. according to that review, Random Combined Read/Write Response Time (Maximum) seems to be more critical on this issue, although i don't know whether this is true or not. as that review shows this random combined response time on intel G1 is nine times longer than P256, and five times than vertex, let's assume G2 gets significant improvement, (write speed doubled than G1 from other review) can we assume G2 improve about 200%(i.e. this number would be half of G1), then it's still way more than double of vertex number and four times of P256 number. but we never know the truth on this regard for G2 from available online resources so far...

for that full format issue, as long as GC or TRIM works, it's not a big issue i guess.

Like I said, I can't read German, but I can read numbers in almost any language:

Quote:
"Eine Windows 7 Bootzeit auf der SSD von nur 17,9 Sekunden spricht eigentlich für sich und verdient sich auch unseren Performance Award für die OCZ Vertex."

Above quote from:
http://www.tweakpc.de/hardware/tests...vertex/s14.php

Translated to English:
Quote:
"TweakPC (Germany)
“The OCZ Vertex shows a marked improvement over the last generation of SSDs. The performance almost makes the mouth water. While reading the economic SSD up to 250 MB/second and write 128k blocks from their maximum of nearly 200 MB/sec. These are truly outstanding values in this test. If you are looking for maximum performance for a system disk, the OCZ Vertex is the SSD you are looking for. A Windows 7 Boot Time on an SSD of just 17.9 seconds speaks for itself and earns our Performance Award for the OCZ Vertex.”"

Translated to English quote above from this page:
http://www.ocztechnology.com/product...ata_ii_2_5-ssd

(Scroll around 55% down the page or search for 'TweakPC (Germany)')

Finally, the testing methodology is found here:
http://www.tweakpc.de/hardware/tests...vertex/s02.php


I may still be wrong, but I am attempting to verify the accuracy of these claims as you can see - I just hope you're doing the same too.

Also, I'm not doubting your numbers (although I can), what I am doubting is that is there any real world basis for 'believing' in those numbers? I agree there is no universal test for everyones needs - but that doesn't negate the fact that benchmarks will always show what the manufacturer or the reviewer wants them to show - not what you think it is.

[vostro1400user]

(Scroll around 55% down the page or search for 'TweakPC (Germany)')

yeah, i referred to this, didn't noticed it is translated from german.

[tilleroftheearth]

vostro1400user,

ah! Always, always go to the source.

[vostro1400user]

"Samsung will be offering support tools and firmware updates for Windows 7 which will be available in the beginning of 2010.

If you wish to be informed as soon as the Windows 7 Firmware Update can be downloaded please register below for our email notification service about general availability."

http://semicon.samsung.de/relaunch/p...are_update.asp

[tilleroftheearth]

lol, yeah, but I read in the summer that Samsung will be offering Win 7 TRIM support in September... lol

I won't sign up for the email notification - I'll just wait for the Samsung user's party that will happen when Samsung eventually does get their act in gear.

[ronan_zj]

lol, just flash 1901Q on ur samsung, it works good.
My friend flashed his Superstupid SSD with 1901Q, and he did feel the boost.