Sign in to follow this  
Kevin OBrien

OCZ Everest 2 Performance Preview Discussion

Recommended Posts

This week at CES we had to opportunity to see the new Indilinx-based Everest 2 reference board in action at the OCZ suite. While it is easy to look at performance numbers on paper, it's entirely different seeing these results in front of you using our same accepted benchmarks. With a quick IOMeter preview, we were able to see the future OCZ Vertex 4 in action working with a 100% random 4K write workload.

OCZ Everest 2 Performance Preview

Share this post


Link to post
Share on other sites

Based on the same IOPs at high end, this SSD does not scale and so the more load you put on it the slower it will get (each process will experience degredation because the IOps are distributed) where the sandforce will have a more consistent performance experience because its overall performance increases as the load increases meaning each individual thread will maintain its IOPs (not exactly but not linear).

Share this post


Link to post
Share on other sites

Not necessarily, if you look at the Octane the performance kind of hit its top end early on but it stayed flat throughout the higher queues.

If the Everest 2 is able to push higher speeds right off the line it would be way more beneficial to end users who would feel more impact from that versus guys who might load the drive down in a heavy multithreaded environment. That and from what we have seen so far the Everest 2 is well above the Sandforce controllers in both read and write speed as is with incompressible data. Scaling well or not at higher queues it is still pushing higher speeds.

Share this post


Link to post
Share on other sites

Very impressive!

Also note the extremly low maximum response time. Some manufacturers increase maximum throughput at the cost of latency by keeping more requests in flight. This doesn't seem to be the case here.

BTW: did they say they used MLC for this demonstration? SLC would be.. unfair :P

Based on the same IOPs at high end, this SSD does not scale

Considering that the performance at QD 3 is miles above what we currrently have, I'd rather say "looks like they managed to work even better at low QDs". And, as TSullivan said, I wouldn't consider flat scaling a bad thing if the overall performance is better than the competition: all tasks are still finished earlier, irregardless of scaling and QD.

MrS

Share this post


Link to post
Share on other sites

Hmmm, we didn't specifically ask about NAND, but I'd assume (I know) it's MLC. They've never made an announcement or Everest configuration that I'm aware of that uses SLC. If they used SLC in such a demonstration, it would be deceptive and bad for business. On the other hand, they do support TLC, maybe that's what it was ;)

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this