Hi all, Ive been looking at a few SSD's and some have cache some dont, does it really matter that much? I know it improves r/w speeds but if i get one without will i even notice? As in will i regret it? I dont think it matters all that much, but i thought id ask as i cant find a definitive answer :/ Thanks guys.
The introduction of cache was predominantly as a way to get around the stuttering issues with the Gen1 SSDs in the Gen2 ones. As Ph4ZeD kind of says (well i'll pretend that he meant this), different Gen2 SSDs use different controllers &, as they've improved over time, some (well the SandForces - can't recall any others(?)) have advanced to the stage where they do not need any cache. Now, that's not to say that additional cache can't have a positive effect on top of this - for example using a decent 6Gb/s raid card with 512MB of cache will show both an improvement in b/ms & r.l. usage vs the same SSDs attached to an on-mobo solution - however this isn't the real reason why it's used on most SSDs.
Do they even sell SSDs anymore with the old SF1500? That chip was the cause of most early ssd problems.
i'm not sure what you mean by "the old SF1500... ...was the cause of most early ssd problems"? The SF1500 is a Gen2 controller that generally outperforms the SF1200 (used in the V2/F100/etc) &, at least on the OCZ front, is used in the Vertex LE (with certain features disabled) & the enterprise level Vertex2 Pro (MLC) & Vertex2 EX (SLC) SSDs. Now, if i'm remembering correctly, the LE was out in Q1 this year (the other two at the end of Q2), whereas the first gen2 SSDs (with cache) were Q2 2009... ...& it's those 2009 SSDs that largely worked around the stuttering problem in the older SSDs by having inbuilt cache (plus improved controllers - though, afaik, still being reliant on cache until the SF controllers came on stream) - & the issue appeared with almost all (i think it may only have been the original ioDrive that didn't) of the nand based SSDs of any brand up to that point. Yeah, so since the SF1500 chip didn't exist in purchasable products until this year... ...well, as said, i'm not sure what you mean by "the old SF1500... ...was the cause of most early ssd problems."
I probably worded that a bit wrong since both chips are infact the same. The difference is in the firmware and setting of the chips. The SF-1500 was intended for enterprise (server) customers. It has smart features built in for power loss and other things the 1200 doesn't have, but these SMART attributes and logging/debug/diagnostic features can also cause some slow down, especially with alot of small writes. The 1200 mass production firmware has been updated alot since its first RC and this has improved small writes alot. I know quite a few people who have owned both a 1500 and a 1200 ssd. All of them had problems with the 1500, stuttering and bad performance. I know its only probably a firmware issue but all of them wouldn't recommend getting an SSD with the SF1500. So what i'm saying basically is that the 1200 is a much improved firmware over the 1500.
Fair enough - yeah, there were some reports of problems with the LE initially, but afaik this has all been resolved with f/w updates - well, they wouldn't be selling the other 2 models successfully to enterprise customers if the issues hadn't been fixed. Still, (1) it's common with lots of cutting edge tech that there will be issues (well, one of Crucial's early C300 f/w updates bricked large no's of drives) & (2) this isn't anything to do with the OP's question about cache as none of the SF's currently have internal cache.
