Odd there aren't multiple capacity choices, 2tb isn't very much space anymore.

toshiba.semicon-storage.com/ap-en/storage/product/internal-specialty/pc/articles/dt02aca7200rpm-series.html

Ah yes, 2 TB 7200 rpm spinner, just what everyone needs for performance critical applications :roll:

Yes thats what we need more SMR drives for non archiving purposes.

ChaitanyaYes thats what we need more SMR drives for non archiving purposes.

Can I have some cancer while we're at it?

This is quite the weird PR piece

This is really weird, the competition is announcing 20TB+ and getting closer and closer to 30TB and toshiba decides to go the opposite direction with a 2TB drive? and SMR no less!? :kookoo:

It's mid 2022 Toshiba: Let's launch a 2TB HDD that's 7200 RPM and 256MB of cache with SATA 6 Gbit/s interface that essentially doesn't set itself apart at all from a Seagate Barracuda has been around for eternity. It's clear there is a lack of effort to make a more reasonably performing HDD today due to product launches like this that illustrate a absolute disconnect from reality. What a time to be alive SATA 6 Gbit/s in mid 2022.

InVasManiIt's mid 2022 Toshiba: Let's launch a 2TB HDD that's 7200 RPM and 256MB of cache with SATA 6 Gbit/s interface that essentially doesn't set itself apart at all from a Seagate Barracuda has been around for eternity. It's clear there is a lack of effort to make a more reasonably performing HDD today due to product launches like this that illustrate a absolute disconnect from reality. What a time to be alive SATA 6 Gbit/s in mid 2022.

"achieving a maximum of 210 MiB/s"

Good luck getting that from an old Seagate Barracuda.

InVasManiWhat a time to be alive SATA 6 Gbit/s in mid 2022.

More than enough for spining rust, can be even considered overkill, not even enterprise 10000+ rpm drives are able to touch the 6gbit limit of sata3.

Aside from expanding storage density HDD innovation has been virtually dead for well over half a decade. I imagine the controller itself is the weakest link of HDD's today relative to NVME followed by cache size and the interface that last point being largely irrelevant until they address the other issues. This device is still woeful for mid 2022 though no matter how you look at it.

trsttteMore than enough for spining rust, can be even considered overkill, not even enterprise 10000+ rpm drives are able to touch the 6gbit limit of sata3.

Sort of yes though I feel the controller and cache size limit it's potential and if those were worked around better the interface issue would come into play further. PrimoCache has shown there is room to block level cache accelerate HDD's to good effect. The big issue is they aren't trying worth a damn to do so all the focus is on NAND innovation which is pretty understandable though storage density is still favorable for spinning rust or optical as well and magnetic tape storage the gap has been narrowing though with spinning rust year over year. In terms of the interface you need to consider the cache performance and controller performance that are technically buffers to the slower spinning storage performance itself. Both the controller's cache structure and configuration and DRAM plays a role and are limited by the interface. SATA 3 will very much limit DRAM look no further than a I-RAM it was crap due to the interface in huge part.

I agree that the sentiment here is that there is nothing outstanding about this drive.

InVasManiAside from expanding storage density HDD innovation has been virtually dead for well over half a decade. I imagine the controller itself is the weakest link of HDD's today relative to NVME followed by cache size and the interface that last point being largely irrelevant until they address the other issues. This device is still woeful for mid 2022 though no matter how you look at it.


Sort of yes though I feel the controller and cache size limit it's potential and if those were worked around better the interface issue would come into play further. PrimoCache has shown there is room to block level cache accelerate HDD's to good effect. The big issue is they aren't trying worth a damn to do so all the focus is on NAND innovation which is pretty understandable though storage density is still favorable for spinning rust or optical as well and magnetic tape storage the gap has been narrowing though with spinning rust year over year.

The limitation has nothing to do with controller or link, there's simply a limit to how fast you can make spinning disks rotate and a seek head, well, seek reliably and effectively. Even the number of platters per disk is a limit difficult to overcome which is why increases in density have moved slower and slower and techniques like shingled recording appeared. As for speed, you have hybrid designs including a "tiny" solid state cache (tiny as couple gigabytes), but that's still not enough to move the needle much when the disk can only spin as fast.

Whenever faster controllers are necessary they'll be ready to go, like they are in SSDs were they are necessary.

The I-RAM was limited by link speed as for the controller we've seen it makes a impact readily in NVME devices you can't tell me it wouldn't effect a HDD likewise. Yes the platter speed is slow and seek speeds cause performance delays, but software like Primo Cache can work around that limitation so why can't the device itself be made to do likewise!!?

What gets in the way of that the processor and the cache itself so the controller and DRAM and the DRAM is limited by interface as is the controller just as it is for NVME.

InVasManiThe I-RAM was limited by link speed as for the controller we've seen it makes a impact readily in NVME devices you can't tell me it wouldn't effect a HDD likewise. Yes the platter speed is slow and seek speeds cause performance delays, but software like Primo Cache can work around that limitation so why can't the device itself be made to do likewise!!?

What gets in the way of that the processor and the cache itself so the controller and DRAM and the DRAM is limited by interface as is the controller just as it is for NVME.

At that point just buy a lot of RAM and make a RAMdisk yourself.

I haven’t used a HDD since 2014 and I don’t regret a thing. SSD’s all the way for performance and longevity.

ExcuseMeWtfAt that point just buy a lot of RAM and make a RAMdisk yourself.

It doesn't require much DRAM to cache lots of storage however. We see that clearly both with NVME caches and with PrimoCache. The bigger issue is 3D stacked NAND is quickly erasing the storage density advantages HDD once easily held over NAND especially taking into account the number that you can physically fit into a server and space occupied doing so and power as well.

I think there are other storage mediums where density is really good that paired with the right interface and sweet spot ratio on DRAM cache buffer along with a strong controller or CPU or what have you could be advantageous. I have to wonder of quantum computers have been used along with high density tape storage in some really peculiar manners to cache accelerate the high density storage in ways that aren't otherwise possible with other technology from a processing standpoint.

I still think this product makes zero sense and don't know what Toshiba was thinking to launch this mid 2022. This product launch should just commit Seppuku it's so late for product with these specs that it feels pretty dead on arrival.

InVasManiIt doesn't require much DRAM to cache lots of storage however.

That depends on needs, working mostly with larger files would need a larger cache

The upside is very minimal beyond about a 1000 : 1 ratio. You might see some upside, but it's inherently quite negligible and really just a overall waste for the type of performance uplift it could or would selectively offer. I'm not sure the reason to that, but seems to hold pretty true of PrimoCache and with NAND in terms of DRAM ratio to it for NVME devices. I wonder if maybe the system timer and polling interrupts is part of the reason. You can't get a system timer lower than 0.5ms and when talking between two devices it would add up to 1ms best scenario. Improvements beyond that is probably largely futile.

Perhaps I'm mistaken about that, but I haven't seen much that indicates helps in any significant manner. There are cases where the CPU can provide a bit of additional upside and I suppose controller like wise by leverage the cache design itself though it is a very small uplift given the size of caches on either today though 5800X3D and EPYC is a slight exception in terms of cache size those are pretty sizable and can make a big impact with caches provided transfer sizes aren't over significant though helps with compressed data to a larger extent.

Based on what I've seen if a bigger cache made any sense NVME devices would be pushing to increase it more than they have, but they've pretty much always adhered to a 1000 : 1 ratio if you've been paying attention to how much DRAM gets paired with given NAND storage capacities. I have yet to see a NVME 1TB with 2GB DRAM or 4GB DRAM and I also don't anticipate that happening with that amount of storage capacity at the same time.

RavenmasterI haven’t used a HDD since 2014 and I don’t regret a thing. SSD’s all the way for performance and longevity.

You must have little data to store or a lot of money to burn, which isn't the case for everyone.

7200RPM feels like the unwanted child.

Noisier and more power use than a 5400RPM but slower than a SSD.

timta2You must have little data to store or a lot of money to burn, which isn't the case for everyone.

1TB SSDs can be had for <100$ and will satisfy storage needs of most users as well.

My new PC has Kioxia Exceria 1TB + older Transcend 370S 512GB and that's plenty for me.

I hope they still produce the 2TB CMR for a long time, better try producing cost efficient 4-8TB CMR.
My 3TB CMR are getting old and still no worthy successor...

If they had improved SMR >1000% where it counts it's possible worth a look.
Same replies from nearly all others I gifted some 4-8TB SMR HDDs from WD and TB... and never use sync on them ;)

Arkz"achieving a maximum of 210 MiB/s"

Good luck getting that from an old Seagate Barracuda.

the old barracudas are considerably better than the new ones

BeerbamI hope they still produce the 2TB CMR for a long time, better try producing cost efficient 4-8TB CMR.
My 3TB CMR are getting old and still no worthy successor...

If they had improved SMR >1000% where it counts it's possible worth a look.
Same replies from nearly all others I gifted some 4-8TB SMR HDDs from WD and TB... and never use sync on them ;)

There's no value for customers on a 2TB SMR drive, be it enterprise (2tb lol) or consumer (low capacity drives have a price floor because of the price of the motor, casing, etc.), the savings here are all pocketed by Toshiba. There's arguments for SMR on very high density drives, not on bottom of the barrel density.

The best thing to do continues to be google the drives extensively before purchase and try to stick with medium density where CMR is still the norm. Also shuck externals if you're savy and adventurous

This has to be for some kind of bottom-of-the-barrel OEM build. Like a $500 computer that can't afford an SSD...