Intel Z68 SRT quich benchmark

[Xelene]

One of the goals I had with my new Media Centre was to test out Intel SRT, I have a larger SSD on my Desktop but whilst it is great, its also a bit time intensive to manage.

My HTPC has a Gigabyte MATX Z68 board, getting SRT working took a bit of effort in BIOS upgrades, Firmware upgrades and Intel driver upgrades.

I tested with a Crucial M4 64GB and WD 2TB drive on the SATA3 ports.

The left hand results are with a bedded in SRT, as you can see it doesnt do much for sustained workloads but really helps with Random reads. The right hand result is the direct bench of the spinning disk after disabling the cache.

My initial thoughts tend towards saying this is a great way of accelerating the typical random read workload you get on a PC, you wont get SSD level performance and its not going to help with sustained stuff like game level loads. The system certainly feels like its on a SSD.

Good tech, seems stable and worth the 60 quid or so for a small M4.

Edit: Interesting thing about small block reads is they start very slow, but as the data is hit again they get cached and the line speed rockets. The driver is really optimised for non sequential workloads.

There is still some concern over drive wear using this. It might be a good way to destroy an SSD quickly.

 

[Xelene]

Thought it would be worth posting my experiences of SRT after 3 months of use.

[/URL]

Dropped the SSD out of cache to check the health of it, still fantastic. Caching hadn't hurt the lifespan at all.

The benchmarks were the same as a new SSD, slightly slower writes + much much faster reads, compared to the benchies of the disk uncached.

I don't have the benches of the SSD in aged cache to show, but it performed much like the RAW disk. The driver had learnt what to cache and what not to cache so overall system snappiness felt like there was a SSD there but you couldn't see it. Reads & Writes for benchmarks felt like the 5200RPM drive - but OS and core games felt like an SSD.

Very good technology IMO.