Performance Comparisons – Client QD Weighted
These results attempt to simplify things by focusing on what really matters – the Queue Depths that folks actually see when using these products. A dimension is eliminated from the previous charts by applying a weighted average to those results. The weights were derived from trace recordings of moderate to heavy workloads, which still ended up running closer to QD=1-2 even on a slower SATA SSD. The intent here is to distill the results into something for those wanting 'just the facts' to grab and go when making their purchasing decisions. Don't be alarmed by the low figures. Remember, these are low queue depths – the place where these SSDs actually operate when in use by those not just running benchmarks all day!
The 750 EVO does surprisingly well here, even at the lower capacities (which have fewer dies to spread the workload). The WD Blue shows the same higher read / slower write trend seen on the previous page, but we were only sampled the 1TB model of that part, so the higher performance must be tempered by the capacities not being on parity. The Intel 600p is the lowest performer on low-QD reads while its higher interface bandwidth and SLC cache helps it beat the pack on writes.
Most of the SATA parts here ride close to interface saturation. The 600p, despite being an NVMe part, is unable to reach its full throughput at lower queues, limiting its lead on the SATA parts.
For those curious how these results pan out in comparison to older / other SSDs, well, here's a few hundred results for your viewing pleasure:
…did I mention we have thoroughly validated our new test suite?
On the page “Performance
On the page “Performance Focus – 750 EVO 250GB” under the first graph it says “Very impressive speeds for the 1TB 960 EVO. […]”.
Clearly that’s wrong 😀
Fixed. Thanks!
Fixed. Thanks!
You’re welcome 😉
You’re welcome 😉
Was Samsung 840 EVO really
Was Samsung 840 EVO really worth Editor’s Choice Award?
840 EVO? Back when it
840 EVO? Back when it launched? Sure. There were issues that were fixed, but could not be discovered at the time of the review.
hey Allyn, is it possible to
hey Allyn, is it possible to include some raid0 SATA devices on your chart? for example samsung 850 pro raid 0 or 960 pro raid 0 to see how it fairs with single drives.
I understand due to raid latency, QD1 performance would drop but since your chart shows average of 1-4QD this would see some improvement in terms of raid, also see how well does SSD caching with intel RST would benefit us over single drive.
Isn’t the 750 EVO EOL now?
Isn’t the 750 EVO EOL now?
I love your write up about
I love your write up about the Latency Percentile. Your storage reviews are by far more realistic with some real engineering behind it. Keep up the great work!
Still waiting on a Storage
Still waiting on a Storage Leader board, like have the stats of all of them on an consistently updated page.
The 500GB 750 Evo is $241,
The 500GB 750 Evo is $241, the 500GB 850 Evo is $170. You’d be a fool to buy the lesser 750 for more than the better 850.
You would be a fool to pay
You would be a fool to pay that, especially since you can get one practically anywhere for ~$145-155 depending on tax/shipping.
Hi Allyn.
Wishing you all a
Hi Allyn.
Wishing you all a happy festive season………………
Samsung and Sandisk(rip)were the only ones to get a grip
on planer TLC.That Ultra 2 result in the write cache test
is really strange.
I remember when you did a comparison a while back I asked
if you could include ultra 2 which was using folding on
each individual die-obviously you were too busy.
Guess something was going on in the background there…..
As to the 750 EVO’s-the 250 and 500 pass my requirements.
1.More than 8000 IOPS read QD1.(must for a boot drive)
2.Write more than 200MB after the cache(cant have it slower
than my spinning rust)
Would be great if this huge
Would be great if this huge chart was searchable and not as an image. I wonder if my Toshiba Toshiba THNSNJ is somewhere there…