For the comparison stats, I compare the rand/seq read/writes vs the best values I can find. Is it fairer to compare vs a single model, instead of the best stats from different models?
I put the seq values first. Change and put the more realistic rand values first?
Your suggestions, comments, and criticisms would be appreciated.
you still work on a guide? haven't seen in posted in ages
Jose Adams
It is stickied.
Isaiah Walker
why don't you just approximate average between official and user benchmarks?
Ethan Phillips
Isn't most of the benefit from SSDs to do with reducing the latency from milliseconds to nanoseconds anyway, not sustained transfer rates? I'd imagine most of the people who read your guide don't say things like "Well I run databases that I know do a lot of 4k random writes at a high queue depth..."
Unrelated: Have you ever considered adding a thing on your website for upcoming new hardware? In the past several months I've seen a lot of people saying they were/are holding off upgrades waiting for Pascal/Polaris/Zen/Broadwell-E/whatever, it'd be nice to have a notification that "such-and-such is scheduled to be released in four months, consider waiting for such-and-such component"
Jaxson Edwards
i see
why does the site use so much javascript?
Elijah Butler
Also your site works like total ass on phone, it's 2016 m8 make a mobile version
Robert Flores
Best SSD should be compared to the best SSD in that same category- Compare m.2 SSDs to m.2, SATA3 to SATA3, PCIe to PCIe.
Looks like your 'compared to best SSD' is comparing SATA3 to PCIe, which is not very good because of the various problems with PCIe SSDs.
Grayson Perry
Because P O O O O
Elijah Collins
do a case fan guide
Mason Hernandez
add MTBF and Warranty, it would be nice it.
Noah Walker
I'd go with the official™ reviews benchmark since they're generally consistant between products.
As for the comparison I don't think it's really necessary to put it in the hover text since speeds are already listed and people will only compare it to others in the same "tier". It doesn't really matter how an 850 evo stacks up to a 950 pro.
David Hall
I need to have a proper source to point to. Much more reliable than approximations, particularly if someone asks: "What is your source?"
Good point. Problem is that latency is not an easy metric to comprehend for those with little/no PC knowledge. I will need to add IOPS too, and that is even less comprehensible to the average Joe.
