Comments on: AS Benchmarks for RealSSD C300 http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/ Learn about Micron's cutting edge innovations in memory technology. Micron's extensive patent holders, world-class scientists and engineers are pathing the way for memory innovation for computing, mobile, server and appliances. Sat, 04 Feb 2012 14:58:52 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Jaundalynn http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-16751 Jaundalynn Fri, 04 Nov 2011 11:24:21 +0000 http://www.micronblogs.com/?p=795#comment-16751 It's great to read something that's both enjoyable and provides pragmatic solutions. It’s great to read something that’s both enjoyable and provides pragmatic solutions.

]]> By: sf http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-3982 sf Mon, 22 Feb 2010 20:20:36 +0000 http://www.micronblogs.com/?p=795#comment-3982 Hi, Not sure if you have read Anandtech's latest SSD state of the nation update. One item that was troubling was the C300's problem with latency. You AS write latency numbers are a bit higher than the competition. It stands out because it is the only area where your drive is slower than one of your competitors. http://www.anandtech.com/storage/showdoc.aspx?i=3747&p=3 In the Anandtech article the worst case scenario was 1.27s for write latency. That is a number that you do not want spreading around. Anand even brings up the dreaded Jmicron word. Anand aludes to the background garbage collection and reorganization of data which are contributing to this problem and that his max latency case might never occur in normal opperation which is indicated by the average write latency. On a side note was your SATA 6Gbps HBA a x1 PCIe card or faster? Anand was using a x1 card and I was interested to see if more can be expected from a card not limited to 500MB/s. Hi, Not sure if you have read Anandtech’s latest SSD state of the nation update. One item that was troubling was the C300′s problem with latency. You AS write latency numbers are a bit higher than the competition. It stands out because it is the only area where your drive is slower than one of your competitors.

http://www.anandtech.com/storage/showdoc.aspx?i=3747&p=3

In the Anandtech article the worst case scenario was 1.27s for write latency. That is a number that you do not want spreading around. Anand even brings up the dreaded Jmicron word.

Anand aludes to the background garbage collection and reorganization of data which are contributing to this problem and that his max latency case might never occur in normal opperation which is indicated by the average write latency.

On a side note was your SATA 6Gbps HBA a x1 PCIe card or faster? Anand was using a x1 card and I was interested to see if more can be expected from a card not limited to 500MB/s.

]]> By: chris http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-2078 chris Fri, 11 Dec 2009 06:17:23 +0000 http://www.micronblogs.com/?p=795#comment-2078 Wow, those numbers are good! Will you ever release a Manual Trim app for OSX. Since OSX doesn't support trim was wondering if you guys planning on releasing a manual trim for your drives. Oh was wondering an estimate for the MSRP and time you'll release these awesome SSDs in the wild? Many thanks! Wow, those numbers are good! Will you ever release a Manual Trim app for OSX. Since OSX doesn’t support trim was wondering if you guys planning on releasing a manual trim for your drives.

Oh was wondering an estimate for the MSRP and time you’ll release these awesome SSDs in the wild?

Many thanks!

]]> By: GullLars http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-2077 GullLars Fri, 11 Dec 2009 05:35:26 +0000 http://www.micronblogs.com/?p=795#comment-2077 Sweet numbers. It would have been nice to see "used"/"sustained" numbers also. Used being after sequentially writing the entire volume 2-3 times, and sustained being directly after writing the volume a some times with small block random IO at high queue depths and leaving about 90% of the LBAs "valid". For referance, the score (627 or 693) is higher than x25-E 32GB at about 550, x25-M 80GB at about 425, kingston V 40GB (an x25-M with only 5 NAND chips, running half the channels) at about 350, and Barefoot 128GB at about 225 points. Still, it's lower than 2 x25-M 80GB in RAID-0 from a motherboard at about 850 points (new) or about 690 points (used), wich can be had at a lower price, and also beats the PCmark Vantage scores. Anyways, RAID does not yet allow TRIM and takes up more ports, pluss higher volume failure rates and more difficult to set up, so it's not a fair comparison. At the time of launch, this will probably be the highest performing consumer SSD on the market. I drool a bit when i think of a few of these on an LSI 9211-i8 HBA (wich can handle over 2GB/s bandwidth and up to 300K IOPS). 6 of the 128GB model in RAID-0 on such a card may deliver 300K 4KB random read IOPS and 240K random write IOPS, and a bandwidth of more than 2GB/s read and over 1GB/s write. Though, the price of such a setup would be right above $2000. Sweet numbers. It would have been nice to see “used”/”sustained” numbers also. Used being after sequentially writing the entire volume 2-3 times, and sustained being directly after writing the volume a some times with small block random IO at high queue depths and leaving about 90% of the LBAs “valid”.

For referance, the score (627 or 693) is higher than x25-E 32GB at about 550, x25-M 80GB at about 425, kingston V 40GB (an x25-M with only 5 NAND chips, running half the channels) at about 350, and Barefoot 128GB at about 225 points.
Still, it’s lower than 2 x25-M 80GB in RAID-0 from a motherboard at about 850 points (new) or about 690 points (used), wich can be had at a lower price, and also beats the PCmark Vantage scores. Anyways, RAID does not yet allow TRIM and takes up more ports, pluss higher volume failure rates and more difficult to set up, so it’s not a fair comparison.

At the time of launch, this will probably be the highest performing consumer SSD on the market.

I drool a bit when i think of a few of these on an LSI 9211-i8 HBA (wich can handle over 2GB/s bandwidth and up to 300K IOPS). 6 of the 128GB model in RAID-0 on such a card may deliver 300K 4KB random read IOPS and 240K random write IOPS, and a bandwidth of more than 2GB/s read and over 1GB/s write. Though, the price of such a setup would be right above $2000.

]]> By: Mike http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-2076 Mike Fri, 11 Dec 2009 05:33:21 +0000 http://www.micronblogs.com/?p=795#comment-2076 Thanks for doing the benchmark. The 64-thread numbers look really great - seems like the C300 does a good job at pipelining random reads/writes. Unfortunately, there's one thing sorely lacking: the write access time. It's roughly around .75ms. This compares to about .079ms on Competitor X's 80GB drive - almost 10x the latency(see http://img24.imageshack.us/img24/2237/bencha.png for a comparison benchmark). Considering one of the main benefits of SSDs is to reduce access time, this is somewhat troubling. Will the write latency issue be fixed in a future firmware version? Aside from this, the C300 looks great so far. Thanks for doing the benchmark. The 64-thread numbers look really great – seems like the C300 does a good job at pipelining random reads/writes.

Unfortunately, there’s one thing sorely lacking: the write access time. It’s roughly around .75ms. This compares to about .079ms on Competitor X’s 80GB drive – almost 10x the latency(see http://img24.imageshack.us/img24/2237/bencha.png for a comparison benchmark). Considering one of the main benefits of SSDs is to reduce access time, this is somewhat troubling.

Will the write latency issue be fixed in a future firmware version? Aside from this, the C300 looks great so far.

]]> By: Elias http://www.micronblogs.com/2009/12/as-benchmarks-for-realssd-c300/comment-page-1/#comment-2065 Elias Thu, 10 Dec 2009 17:54:42 +0000 http://www.micronblogs.com/?p=795#comment-2065 Hi! when it will be available to oem or consumers-in january/february/march? Hi! when it will be available to oem or consumers-in january/february/march?

]]>