- Dec 29, 2009
- 26
- 0
- 0
Per Anand's March 2009 tremendous article (with followup in Aug) ssd anthology, which IMHO completely redirected and accellerated the course of SSD technoligical deployment, which everyone has significantly benefited from (Thanks Anand!), thus creating the distinction between Gen1 (jmicron versions) and Gen2 (indilinx), I respectfully ask that we carry his work forward so the average joe (me) can understand how best to configure ssd raid 0 under win 64-bit and win 7 64-bit.
Again, thank you Anand. You single-handedly brought due diligence to bear on SSD oem's. without you, there is no doubt in my mind that we would not see the meta-leep in ssd latency reduction for at least 2 years, instead of the 4 months that you caused to happen.
Ok, on with the question(s).
First, it appears that two (2) vertex ssd's in 4kb ntfs file size actually outperform four (4) vertex ssd's in 4kb ntfs file size, but that 2 vertext ssd's perform much slower in larger (e.g. 256kb, 512kb ntfs file sizes). Does everyone agree with this assertion?
If not, lets review the data to gain concensus.
If so, the next question is, since ssd's read and write in 4kb page size, but can not erase except in 128kb block size (32 pages at a time), is it true that the best performance (MB/s) ratio between OS NTFS file size and ssd is 128kb OS NTFS file size, e.g. 1:1 ratio.
Or, is it better to have the OS NTFS file size set at 4kb?
And, (assuming i am not way off base on the above), does the SSD take 32 pages to write one OS ntfs file that is 128kb in size.
I'll stop here, to see what happens, and to see if i am "on target" or way off base, or somewhere inbetween...
Thanks to all for your considerations and feedback...
And again, thank you Anand. we would'nt be having this conversation without your strength, courage and integrity...
Jeff
Mich
Again, thank you Anand. You single-handedly brought due diligence to bear on SSD oem's. without you, there is no doubt in my mind that we would not see the meta-leep in ssd latency reduction for at least 2 years, instead of the 4 months that you caused to happen.
Ok, on with the question(s).
First, it appears that two (2) vertex ssd's in 4kb ntfs file size actually outperform four (4) vertex ssd's in 4kb ntfs file size, but that 2 vertext ssd's perform much slower in larger (e.g. 256kb, 512kb ntfs file sizes). Does everyone agree with this assertion?
If not, lets review the data to gain concensus.
If so, the next question is, since ssd's read and write in 4kb page size, but can not erase except in 128kb block size (32 pages at a time), is it true that the best performance (MB/s) ratio between OS NTFS file size and ssd is 128kb OS NTFS file size, e.g. 1:1 ratio.
Or, is it better to have the OS NTFS file size set at 4kb?
And, (assuming i am not way off base on the above), does the SSD take 32 pages to write one OS ntfs file that is 128kb in size.
I'll stop here, to see what happens, and to see if i am "on target" or way off base, or somewhere inbetween...
Thanks to all for your considerations and feedback...
And again, thank you Anand. we would'nt be having this conversation without your strength, courage and integrity...
Jeff
Mich