Corsair Neutron GTX 240 GB SSD

[W1zzard]

To read this review go to: http://www.techpowerup.com/reviews/C...on_GTX_240_GB/

[Delta6326]

Meh not very good drive. The all new Sammy 840 PRO rocks!

I can't believe that at Raid 0 it only gets 105%

[Completely Bonkers]

"GTX". Does the marketing genius that thought up this original name think they deserve a bonus or a severance cheque?

[EarthDog]

Did I miss the synthetic tests? Where is ATTO? CDM? AS SSD? Obviously that isnt the entire story, more like a small piece, but still...

I had to have missed it?

[tacosRcool]

If the $/GB was like $0.74 then I would consider this. OCZ may have higher failure rates, but they have the price

[W1zzard]

Originally Posted by EarthDog

Did I miss the synthetic tests? Where is ATTO? CDM? AS SSD? Obviously that isnt the entire story, more like a small piece, but still...

I had to have missed it?

We don't do synthetics in our benchmarks. Check any other random review site for those. I think real life performance testing with real applications, not traces, will give you a more realistic idea what to expect. Infinitely fast storage will not give you zero application runtime. Do you buy a graphics card based on the fillrate testing?

I am debating adding some synthetic testing to our next SSD bench suite, but the results will not be part of the performance rating. Also always looking for feedback on what additional tests we could add. I find it kinda difficult to identify programs nowadays that really need tons of IO, other than what we already have

[james888]

Originally Posted by W1zzard

Check any other random review site for those. I think real life performance testing with real applications, not traces, will give you a more realistic idea what to expect. Infinitely fast storage will not give you zero application runtime. Do you buy a graphics card based on the fillrate testing?

Why tpu rocks

[Zakin]

Very odd the drive is performing so middle ground here, read about four other reviews showing these things absolutely destroying in latencies for sure, as well as read/writes were right up there with the Vertex, and the Raid 0 absolutely destroyed. I suppose for some reason it isn't translating as well on these benches.

[W1zzard]

Originally Posted by Zakin

Very odd the drive is performing so middle ground here, read about four other reviews showing these things absolutely destroying in latencies for sure, as well as read/writes were right up there with the Vertex, and the Raid 0 absolutely destroyed. I suppose for some reason it isn't translating as well on these benches.

i assume those other reviews were using synthetic benchmarks or traces only?

[Ghost]

Originally Posted by W1zzard

We don't do synthetics in our benchmarks. Check any other random review site for those. I think real life performance testing with real applications, not traces, will give you a more realistic idea what to expect. Infinitely fast storage will not give you zero application runtime. Do you buy a graphics card based on the fillrate testing?

I am debating adding some synthetic testing to our next SSD bench suite, but the results will not be part of the performance rating. Also always looking for feedback on what additional tests we could add. I find it kinda difficult to identify programs nowadays that really need tons of IO, other than what we already have

That's what I wanted to suggest.

Having AS SSD and CDM (both compressible and incompressible) results wouldn't hurt. They can show what to expect when dealing with certain type of data. Also can be used to challenge manufacturer's specifications.

[EarthDog]

Originally Posted by W1zzard

We don't do synthetics in our benchmarks. Check any other random review site for those. I think real life performance testing with real applications, not traces, will give you a more realistic idea what to expect. Infinitely fast storage will not give you zero application runtime. Do you buy a graphics card based on the fillrate testing?

I am debating adding some synthetic testing to our next SSD bench suite, but the results will not be part of the performance rating. Also always looking for feedback on what additional tests we could add. I find it kinda difficult to identify programs nowadays that really need tons of IO, other than what we already have

No but a GPU's rated specs arent based off of fill rates, however like an SSD's specs are based off of ATTO results. You also run synthetic benchmarks on GPU's...last I checked I dont play 3DMark11... same concept here Wizz. Just adds another face to an already good testing suite.

Wouldnt it be nice to reach the specs they list if only for giggles?

As far as suggestions, ATTO, AS SSD, CDM, plus IOMeter (you can customize the load in IOMeter...).

[washd123]

why do you put "needs SATA III to perform the best" as a con? That's more of an issue with the interface itself not the drive. Also even if it does need it to perform the best, is the performance of SATA II that bad? I don't get how it's a con

[N!ck]

IMHO TPU SSD tests are the best because they don't include useless synthetic tests, real Aps FTW! Raid test are also nice - cause a lot of people raid small ssd's.

[MxPhenom 216]

IDK why people are hating on these drives. These and the non GTX drives are great. Reliable and fast. And don't use sand force, thank god.

Originally Posted by washd123

why do you put "needs SATA III to perform the best" as a con? That's more of an issue with the interface itself not the drive. Also even if it does need it to perform the best, is the performance of SATA II that bad? I don't get how it's a con

He has put that for every drive he has reviewed. Its a stand procedure. And its true, it needs SATA III to perform the best.

[Kreij]

Originally Posted by W1zzard

I find it kinda difficult to identify programs nowadays that really need tons of IO, other than what we already have

Since most programs do not require tons of IO, adding non-mainstream programs that are IO heavy would be kind of superfluous and the same as synthetics, no?

Unless you want to try to simulate high use server load (ie. database, VM, RDS) I don't really see the need.

Just my 2 cents.