Review Crucial T705 2TB SSD Review: The Fastest SSD on The Planet

I'm kind of curious. Who buys these top end Gen5 drives?
A top end Gen4 is significantly cheaper and is only microseconds slower at loading Windows/Games.
 
  • Like
Reactions: PEnns
'The 4TB option is especially nice as it’s been challenging to find that much flash in the very fastest drives. This drive is certainly one of those, capable of reaching up to 14,500 / 12,700 MB/s for sequential reads and writes and up to 1,550K / 1,800K random read and write IOPS. '

Might just be the way I'm reading it, but you mention the 4TB and then go on to describe it using the speeds of the 2TB version.

Sorry, I'm being pedantic.
 
I'm kind of curious. Who buys these top end Gen5 drives?
A top end Gen4 is significantly cheaper and is only microseconds slower at loading Windows/Games.
Many users are moving large files back and forth, or are manipulating them, and every second saved can make a difference. But for most of us gamers you're right, the experienced delta and cost thereof isn't worth it. Having the "fastest" is an early adopter tax for those with deep pockets that the manufacturer is hoping will help them recoop their development costs, aka the price of developing, manufacturing, and offering the "fastest".
 
I'm kind of curious. Who buys these top end Gen5 drives?
A top end Gen4 is significantly cheaper and is only microseconds slower at loading Windows/Games.

People with deep pockets and people who chase after specs. The only people that benefit from these drives at the moment are those with workloads where it benefits them or outright requires them. On other words, you use it for work where every second matters, your company has more money than sense, or some combination of both.
 
Their pricing with/without heatsink seems a bit random. Assuming they use the same heat sink for the different models, you would expect the same price increase to include it. But it's not. It's $20 extra for the 1TB and $40 extra for 2TB and 4TB according to the article pricing. I checked Crucial's site and it lists the non-heatsink 4TB for $714 which is only $16 extra for the heat sink. WTF?
 
I feel like we should have some benchmarks of Optane U.2 drives included with these results 😀
If rotated so that the graphs’ bars were oriented vertically, the latency benchmarks would resemble steep cliffs with the inclusion of Intel Optane SSDs—any generation of your choice.

The NAND IOps nowadays, though, are impressive. 1.5 MIOps is what the P5800X SSDs can do (but over 5.0 MIOps if you have a random I/O workload that can take advantage of 512-bytes sectors).
 
  • Like
Reactions: Albert.Thomas
Ehhh, I get why you didn't include more Gen5 drives. If I recall correctly, the T700 swept virtually every category(?), so why include anything other than the previous champ... but it would have been nice to see at least one other Gen5 drive that wasn't Corsair. And yeah, I can go back and look at other reviews to see those, but then I have to go back and look at other reviews.

It just feels weird. Just a nitpick. I always appreciate Shane's reviews and Jarred's benchmarking.

next day edit: oh my gog, I don't even know what I'm talking about. I was running a fever yesterday, and only like 11% of me remembers writing this. Sorry... I'm so embarrassed.
 
Last edited:
At some point, they are going to require active cooling. My 4TB Lexar Gen 4 runs warm even with its good heatsink. I no longer buy i9 CPUs (i7 is good enough) as they run too hot for a nice, quiet and easy to install air cooler. The manufacturers are, in essence, overclocking their products to get these high results. I don't bother anymore.
 
  • Like
Reactions: PEnns
Probably yet another benchmark optimized drive.

Write to that drive from start to finish with an actual dataset, let it heat up and report back to us once you have written it twice.

I'm willing to bet it falls down from that speed before you have written 1Tb to it, probably even before you reached 500Gb.

Now write down that speed for the actual speed, probably not much more than 200Mbps or even half of that

Because if you are just running a benchmark to come to these numbers, like everyone else does... you will never measure its real world performance.
 
  • Like
Reactions: PEnns
'The 4TB option is especially nice as it’s been challenging to find that much flash in the very fastest drives. This drive is certainly one of those, capable of reaching up to 14,500 / 12,700 MB/s for sequential reads and writes and up to 1,550K / 1,800K random read and write IOPS. '

Might just be the way I'm reading it, but you mention the 4TB and then go on to describe it using the speeds of the 2TB version.

Sorry, I'm being pedantic.
It will never sustain those speeds outside of a benchmark.

Reviewers need to write the drive from start to finish with an actual dataset, because even the slightest bit of compression is going to skew these benchmark numbers in favor of the drive.

I'm willing to bet it falls down to sub 200Mbps when I write to it from start to finish.
 
Sounds great. But...what did Crazy Eddy say (those who were born before the 1990s will remember him 😉) ?

Those Prices Are Insane!!
Nah. It's not CRAZY. I'll wait for a drive with similar performance, costs less, and doesn't run as hot.
Sounds great. But...what did Crazy Eddy say (those who were born before the 1990s will remember him 😉) ?

Those Prices Are Insane!!
 
I'll wait for lower price, lower power usage as I don't like the idea of pushing a chip to the mfg specs. Most of us don't need this and wouldn't notice the speed difference. It is simply the idea of having the best that excites some people.
 
Probably yet another benchmark optimized drive.

Write to that drive from start to finish with an actual dataset, let it heat up and report back to us once you have written it twice.

I'm willing to bet it falls down from that speed before you have written 1Tb to it, probably even before you reached 500Gb.

Now write down that speed for the actual speed, probably not much more than 200Mbps or even half of that

Because if you are just running a benchmark to come to these numbers, like everyone else does... you will never measure its real world performance.
You don't have to speculate, sustained performance results are included in this review. Write speeds drop to 4GB/s after pSLC is exhausted, and if you keep going it will periodically drop to 1.2 GB/s when the drive has to do garbage collection. It's a bit of a stretch to call those the "real" performance though, given you only hit the former or latter values when you're writing hundreds of GBs or multiple TBs at a time, respectively. Which most people will rarely if ever do. If you are routinely writing TBs at a time, I'd argue you shouldn't be buying a consumer drive like this in the first place.

Edit: the write thresholds above are for an empty drive, they will decrease if you're starting from a drive that's already part full.
 
Last edited:
  • Like
Reactions: AgentBirdnest
If you are routinely writing TBs at a time, I'd argue you shouldn't be buying a consumer drive like this in the first place.
I’ve been periodically scouring the usual SSD manufacturer websites for press releases, product catalogs, and documentation. I don’t believe there are any PCIe 5.0 SSDs that can sustain more than around 6 GB/s—enterprise SSDs included.

Anyone seeking insane storage bandwidth would have to settle for a PCIe switch/RAID AIC and enough drives to saturate the connection. There aren’t any that’ll do PCIe 5.0, so in short it’s pretty much unobtanium. ¯\_(ツ)_/¯
 
  • Like
Reactions: TJ Hooker