facebook rss twitter

OCZ Vertex 4 SSDs receive BIG boost with firmware v1.4RC

by Alistair Lowe on 8 May 2012, 11:22

Tags: OCZ (NASDAQ:OCZ)

Quick Link: HEXUS.net/qabgcz

Add to My Vault: x

When we reviewed OCZ's first Indilinx-powered Vertex SDD, the Vertex 4, back in April, we applauded OCZ for its great job in moving away from SandForce controllers, whilst producing an heir true to the Vertex-series title. Our reviews found that synthetically, the Vertex 4 generally came out on-top, featuring great IO performance, however, was a little weak when it came real-world benchmarks with misaligned data; variants with less storage also featured inferior write speeds, taking some of the competitive edge away from affordable sub-256GB offerings.

However, an announcement yesterday from OCZ of a new firmware release candidate, V1.4RC, demonstrates some of the advantages in utilising home-grown controllers, with quick and closely managed firmware development cycles. The new firmware claims a significant boost in speeds across the range, with the most notable a 210 per cent increase in write performance for the 128GB Vertex 4 variant.

128GB 256GB 512GB
Previous Specs New Specs Previous Specs New Specs Previous Specs New Specs
Max Read 535MB/s 550MB/s 535MB/s 550MB/s 535MB/s 550MB/s
Max Write 200MB/s 420MB/s 380MB/s 465MB/s 475MB/s 475MB/s

Of course it's not all about sequential read and write speeds, however. Thankfully, the storage enthusiasts over at StorageReview have been playing around with v1.4RC on a 512GB Vertex 4 and have posted initial benchmark results. Findings show that whilst synthetic benchmarks remain fairly consistent, IO performance in IOMETER, which focuses on simulating real-world load scenarios, sees an increase anywhere from 16 right up to 75 per cent.

Likewise, real-world benchmarks involving measurement of throughput show performance gains from 15 per cent under a HTPC-like workload, with gains of up to 61 per cent when gaming, which has a heavier demand on drive read performance. In all real-world instances, throughput, IOps and average latency showed significant increases in performance.

Whilst we await the final release of firmware V1.4, perhaps OCZ should give itself a pat on the back for a job well done.



HEXUS Forums :: 9 Comments

Login with Forum Account

Don't have an account? Register today!
Wasn't it discovered that infact the controller isn't home grown, its a Marvel with heavily customised firmware?
Yeah it was, which means there is no excuse for it to be any worse than other Marvell drives like the m4, and in some benches it was way off the pace until now. Rushed out the door again and fixed later by the looks of it. If the differentiator is the quality of your firmware you need to get it right on release, not months later!
Biscuit
Wasn't it discovered that infact the controller isn't home grown, its a Marvel with heavily customised firmware?

Yes, but there's a lot of emphasis placed on the firmware that runs on-top and simply by owning the firmware development you have the same benefits as if the chip was made in-house.
Seeing as this is a Release Candidate, does that mean there may be more gains in performance for the final release?
Scribe
Yes, but there's a lot of emphasis placed on the firmware that runs on-top and simply by owning the firmware development you have the same benefits as if the chip was made in-house.

But how does that differ to the amount of work/control that goes into sandforce controller based drives?