LinkBack URL
About LinkBacksRead more.A new metric for SSD speed measurement.
Read more.A new metric for SSD speed measurement.
A couple more generations from now these things will be affordable and capacious enough to overtake mechanical HDDs. I can't wait to be able to afford stick one of these in my PS3.
what gains would you get from sticking one in your PS3?
Hmmm 240GB. Tempting. I have started to use SSD for boot drives in my home servers, its well worth it.
Did you know that the EVE Online cluster is all SSD?
□ΞVΞ□
About 15cm gain in epeen sizeOriginally Posted by matty-hodgson
It's 4.98*
I really hope vrpm doesn't catch on. It makes no sense.
PHP Code:$s = new signature();
$s->sarcasm()->intellect()->font('Courier New')->display();
especially once SSD completely takes over HDD, if that ever happens..
I didn't know that the EVE Online cluster is all SSD; I did however know that they have multiple RamSans:
http://www.superssd.com/products_sub.htm
which whilst SSD, aren't SSD in the classic sense
How would you rather compare the performance of two components performing the same task but based on hugely differing technologies? There will inevitably be a period of time where low capacity SSDs and higher capacity HDDs share price points and consumers will have to choose between them. If the industry can find a reliable measure of relative performance across the 2 technologies then (speaking as a consumer) I'm massively in favour.for the sake of pragmatism, let's hope this quasi-standard doesn't stick.
I'm not saying that vRPM is the right measure, of course, but if it can be verified by other manufacturers as giving an accurate reflection of performance (i.e. is a 40,000 vRPM SSD really 8 times faster than a 5000 RPM HDD), and an independant body can be found to maintain the software (so manufacturers can't just fake their results), then I see no "pragmatic" reason why vRPM shouldn't become an industry standard...
Because RPM is not directly linked to data rate. As RPM remains at 7200 but platter density increases, so too does data rate. vRPM is therefore bollocks as a metric.
They should measure burst, sequential and random read/write access times and just state those. Hell, for most people sequential read/write figures would be enough - only SMEs and up will really care that much about anything else.
PHP Code:$s = new signature();
$s->sarcasm()->intellect()->font('Courier New')->display();
Clunk (13-01-2009)
Errr.... how about using some properly reproducible numbers in FIXED UNITS. Maybe like MB/s?! For instance?! I mean, how hard is that to do? vRPM is just retarded, as it would depend on what disk you were spinning. I'm sure I wouldn't want to buy an SSD that had a vRPM of 40,000 with respect to a 40 GB HDD from 9 years ago.
I can see plenty of pragmatic reasons why vRPM shouldn't become industry standard.
Right, I'm off to re-label the speedo on my car to use velocities defined in tree-heights per second.
Complete and utter bollocks indeed.
I prefer to measure the speed of my SSDs in smoked mackerels. This is a much better way because they are rich in protein and high in Omega 3 oils which let the data slip through faster.
Before long, the entire industry will be measuring in SM/sec, so watch out for the benchmark.
You owe it to yourself to click here really.
Keva161 (13-01-2009)
Hahahaha
From reading the white paper, the vRPM benching software *does* measure burst, sequential and random data rates - then it converts that into a relative performance metric (alledgedly weighted by real world usage models).
I'd've thought most consumers would like a single, easily comparable figure: suppose comparing 2 disks, A has a much better burst write speed, but B has better sequential write and random read times. Which is going to provide better performance? vRPM is not a completely artificial measure because it's based on actual benchmarked performance: it's just put into one easily comparable metric. People like simple comparisons (look at AMDs PR on Athlon CPUs, v.s. straight clock speeds) and giving them 3 - 5 different measures of performance to compare would put off a lot of people. vRPM may be a dumb name for it, but as a concept I'm definitely in favour...
Parts of games stored to the hard drive would benefit from reduced load times.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote
