The power efficiency numbers look almost too good to be true, so I'm working to verify that and the results in Bench may be updated before the review is ready.
That's just... WOW. Trading blows with the Samsung 970 EVO Plus while showing those kinds of efficiency numbers (assuming they're accurate) is really impressive.
Maybe SK Hynix focusing on building SSDs to be used as original equipment in laptops WITHOUT also having to chase peak performance figures for the retail market has allowed them to design a supremely efficient drive that still has pretty good performance.
The drive also seems to idle at pretty low power without compromising on idle wake-up time.
Really curious to read the review and see if there's any details to be published that can shed some light into how they've achieved all of this... because this looks like a great drive to be used in a road-warrior focused laptop.
Not that Samsung, Crucial/Micron, or WD/SanDisk are even remotely absent from the OEM laptop space... but they have also been selling into the retail channel the whole time and have had to split their focus in controller, firmware, NAND, etc design and development to hit big numbers that seem to matter so much in the retail space.
Yeah very impressive, but still useless for me. I need a 2TB SSD in that form. So the 970 Evo Plus is still the only option for me. If I want a low capacity SSD for the OS, I will buy a 970 Pro instead.
Yeah, I checked. Theres nothing better actually, if you care about speed at an still tolerable price. Others might be better in some irrelevant benchmarks, they might be a lot cheaper, but in real life usage the 970 Evo Plus is still all around the best one for the price. Normally I only buy 9x0 Pro, but they are too expensive at those sizes and theres no 2 TB version anyway.
I don't think layer count directly affects efficiency in any big way. Higher per-die capacity leads to fewer dies per drive which can improve efficiency, provided performance doesn't suffer. Newer generations of NAND also tend to use lower IO voltages. I expect the P31 results will turn out to be the combination of several factors, for both NAND and the controller.
I've been testing power with Quarch equipment since the end of 2017, with an upgrade in spring 2019 that mainly affected idle power measurement. Before getting the Quarch stuff, we used an Extech multimeter for years, with some custom wiring to measure current for SATA drives and an Adex riser for PCIe drives.
Ah, OK, that makes sense. Increased layers also comes along w/ new generations of NAND, so it's not directly comparable. And the controller, yes: I've only recently realized that SSD controllers are dual-core vs tri-core vs penta-core. It makes sense re: power consumption.
That's great to hear; performance looks solid (so far) from a $0.15/GB NVMe SSD.
Oh, yes. I remember the methodology about the Quarch equipment! I see the button now (😅) for "Energy Usage" at the bottom of the ATSB graphs in the reviews. Energy is the right word, but my mind looked for "power consumption".
The power consumption data is quite thorough; looking at the latest 870 QVO review, seeing power vs read/write mix is pretty interesting, esp. comparing 1 TB vs 4 TB. The difference is there, but relatively small, all things considered.
When can we expect the review to be available? I am eyeing ADATA XPG SX8200 Pro at this point because of its performance/power efficiency but if this SSD is upto it, then I can go for this.
I'm trying to get it published later this week. I have a few questions I want to discuss with Hynix first if possible, but we probably won't delay the review for that alone. Once I have a complete set of believable results (hopefully by the end of the day), I'll write up the review and publish it when it's ready.
I just saw the preliminary Benchmarks. Except for 'Idle Power Consumption metric', the overall results are vastly better than ADATA one. Eagerly waiting for your Final words and Conclusion :)
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
20 Comments
Back to Article
Chaitanya - Tuesday, August 18, 2020 - link
Price looks quite reasonable, I would like to see some reviews of this SSD.Billy Tallis - Tuesday, August 18, 2020 - link
Initial, partial test results are in Bench: https://www.anandtech.com/bench/product/2627?vs=23...The power efficiency numbers look almost too good to be true, so I'm working to verify that and the results in Bench may be updated before the review is ready.
MrCommunistGen - Tuesday, August 18, 2020 - link
That's just... WOW. Trading blows with the Samsung 970 EVO Plus while showing those kinds of efficiency numbers (assuming they're accurate) is really impressive.Maybe SK Hynix focusing on building SSDs to be used as original equipment in laptops WITHOUT also having to chase peak performance figures for the retail market has allowed them to design a supremely efficient drive that still has pretty good performance.
The drive also seems to idle at pretty low power without compromising on idle wake-up time.
Really curious to read the review and see if there's any details to be published that can shed some light into how they've achieved all of this... because this looks like a great drive to be used in a road-warrior focused laptop.
MrCommunistGen - Tuesday, August 18, 2020 - link
Not that Samsung, Crucial/Micron, or WD/SanDisk are even remotely absent from the OEM laptop space... but they have also been selling into the retail channel the whole time and have had to split their focus in controller, firmware, NAND, etc design and development to hit big numbers that seem to matter so much in the retail space.Beaver M. - Tuesday, August 18, 2020 - link
Yeah very impressive, but still useless for me.I need a 2TB SSD in that form. So the 970 Evo Plus is still the only option for me.
If I want a low capacity SSD for the OS, I will buy a 970 Pro instead.
eek2121 - Tuesday, August 18, 2020 - link
You can get NVME SSDs up to 8TB in size. There are many 2TB drives that are cheaper and faster than Samsung drives.Beaver M. - Tuesday, August 18, 2020 - link
Yeah, I checked. Theres nothing better actually, if you care about speed at an still tolerable price.Others might be better in some irrelevant benchmarks, they might be a lot cheaper, but in real life usage the 970 Evo Plus is still all around the best one for the price.
Normally I only buy 9x0 Pro, but they are too expensive at those sizes and theres no 2 TB version anyway.
Chaitanya - Tuesday, August 18, 2020 - link
That is mighty impressive, now will be waiting for full review.ikjadoon - Tuesday, August 18, 2020 - link
>The power efficiency numbers look almost too good to be true,IIRC, increased layers increases power efficiency.
Very happy to see power consumption added! Is this new?!
Billy Tallis - Tuesday, August 18, 2020 - link
I don't think layer count directly affects efficiency in any big way. Higher per-die capacity leads to fewer dies per drive which can improve efficiency, provided performance doesn't suffer. Newer generations of NAND also tend to use lower IO voltages. I expect the P31 results will turn out to be the combination of several factors, for both NAND and the controller.I've been testing power with Quarch equipment since the end of 2017, with an upgrade in spring 2019 that mainly affected idle power measurement. Before getting the Quarch stuff, we used an Extech multimeter for years, with some custom wiring to measure current for SATA drives and an Adex riser for PCIe drives.
ikjadoon - Tuesday, August 18, 2020 - link
Ah, OK, that makes sense. Increased layers also comes along w/ new generations of NAND, so it's not directly comparable. And the controller, yes: I've only recently realized that SSD controllers are dual-core vs tri-core vs penta-core. It makes sense re: power consumption.That's great to hear; performance looks solid (so far) from a $0.15/GB NVMe SSD.
Oh, yes. I remember the methodology about the Quarch equipment! I see the button now (😅) for "Energy Usage" at the bottom of the ATSB graphs in the reviews. Energy is the right word, but my mind looked for "power consumption".
The power consumption data is quite thorough; looking at the latest 870 QVO review, seeing power vs read/write mix is pretty interesting, esp. comparing 1 TB vs 4 TB. The difference is there, but relatively small, all things considered.
ikjadoon - Tuesday, August 18, 2020 - link
*edit: energy consumption!Spunjji - Wednesday, August 19, 2020 - link
Holy cow! Well, fingers crossed those numbers are accurate, because if so my next SSD has chosen itself.Srikzquest - Tuesday, August 18, 2020 - link
When can we expect the review to be available? I am eyeing ADATA XPG SX8200 Pro at this point because of its performance/power efficiency but if this SSD is upto it, then I can go for this.Billy Tallis - Tuesday, August 18, 2020 - link
I'm trying to get it published later this week. I have a few questions I want to discuss with Hynix first if possible, but we probably won't delay the review for that alone. Once I have a complete set of believable results (hopefully by the end of the day), I'll write up the review and publish it when it's ready.Srikzquest - Tuesday, August 18, 2020 - link
Thank you for the update Billy.Srikzquest - Tuesday, August 18, 2020 - link
I just saw the preliminary Benchmarks. Except for 'Idle Power Consumption metric', the overall results are vastly better than ADATA one. Eagerly waiting for your Final words and Conclusion :)Srikzquest - Tuesday, August 25, 2020 - link
Hi @Billy, any update on the review?Exodite - Tuesday, August 18, 2020 - link
Eagerly awaiting the 2TB+ prices, as this is a market were competition is sorely needed.WD_Stevens - Wednesday, August 19, 2020 - link
Is there any way to buy this except Amazon? They don't send anything to Australia anymore.