Discussion in 'Reviews' started by W1zzard, Jul 20, 2010.
To read this review go to: http://www.techpowerup.com/reviews/OCZ/Agility_2_120_GB/
Thank you for your Review W1z. Finally there are some SSD reviews here on TPU. 8)
One question though: Is it possible that you mixed up the 4K random write and 4K random read speed?
Usually in Iometer the SF-based SSDs score around 160-170MB/s in WRITE performance (with properly aligned writes at 4K boundaries) with low queue depth (read: 1 outstanding IO) and around 190-200MB/s with high queue depth (read: 32 outstanding IOs) - at least with the standard Iometer version which writes highly compressible data (read: all zeros) thanks to SandForces DuraWrite. With complete random data this should drop down to around 120MB/s. For some reason the random read speed for small blocks is somewhat worse than the write speed on SF-based drives and usually comes in at 50-60MB/s for 4K random reads. It looks like that this is the exact opposite of what you are measuring?! Can you please verify that?
Nice review W1zz. Now we just need a TPU ssd bios library
The only constructive criticism I can give to you is drivers and southbridge. I have read other SSD reviews (such as techreport) and sometimes there are issues from MS AHCI drivers or vise versa.
To get a really complete SSD review is a ton of work, as I would probably do a run of benchmarks with an Intel SB(ICHR10) with Intel and MS AHCI drivers AND a run of benchmarks using AMD's SB850 Sata6 with MS and AMD AHCI drivers. Then all users have some valid information from the testing and which driver works better for certain tasks.
As for me I have an AMD SB750 so all those benchmarks I have to take with a grain of salt.
Hope my feedback helps for future reviews W1zz.
Keep up the great work!!
that's too much work. i ended up using ahci because that's what the "standard" is and if an ssd fails with ahci people simply shouldnt buy it and go for a product that works. if there is an overwhelming response to use different drivers/platform, no problem i can change that
this ocz ssd is faster than the intel M G2 sdd, no?
For anyone looking at purchasing this drive, tigerdirect.com has it on sale right now and if you're willing to jump thru a few hoops (mail-in rebates, and bing cashback) you can snag this drive for $197!
Great review as always W1zzard. Love to see all the (acutal) real-world benchmarks. It seems like most other SSD reviews focus on IOMeter and HDTach/HDTune for synthetic benchmarks and then for a "real world" benchmark they throw in something like PCMark.
I agree testing each new drive on every chipset / driver combination would be a ton of work, it might be better to take the fastest drive you can get your hands on and run it on each chipset/driver combo to see if there's a bottleneck anywhere. Although that wouldn't really make very good material for a review, more like an experiment I guess.
I know in my experience (SB750) everyone says "use the MS AHCI driver so you have TRIM" but the AMD one is way faster - about 15 seconds faster booting and much better sequential read performance. Plus my drive doesn't have TRIM anyway.
Anyway, can't wait to see more SSD benchmarks on TPU.
@mtosev - I think they are pretty comparable, the agility 2 is faster for some tasks and the X25-M is faster for others. If you google "sandforce vs intel" the first few reviews are helpful.
It would be nice to see benchies on an AMD SB850 though. Doesn't Omega have a Crosshair IV lying around??
You have used AHCI driver, i just wanted to ask is AHCI part of H55 chips there are no Raid functions on that H55 and only ICH with R did support AHCI?
AHCI is needed for NCQ for those that don't know that, but still can't get much more faster than that.
AHCI has nothing to do with raid, and yes it's officially supported on H55 (which is what my test board uses)
Separate names with a comma.