The Intel Optane SSD 800p (58GB & 118GB) Review: Almost The Right Size
by Billy Tallis on March 8, 2018 5:15 PM ESTIntel's first Optane products hit the market almost a year ago, putting the much-awaited 3D XPoint memory in the hands of consumers. Today, Intel broadens that family with the Optane SSD 800p, pushing the Optane brand closer to the mainstream.
The new Optane SSD 800p is an M.2 NVMe SSD using Intel's 3D XPoint memory instead of flash memory. The 800p is based on the same hardware platform as last year's Optane Memory M.2 drive, which was intended primarily for caching purposes (but could also be used as a boot drive with a sufficiently small operating system). That means the 800p uses a PCIe 3 x2 link and Intel's first-generation 3D XPoint memory—but more of it, with usable capacities of 58GB and 118GB compared to just 16GB and 32GB from last year's Optane Memory. The PCB layout has been tweaked and the sticker on the drive no longer has a foil layer to act as a heatspreader, but the most significant design changes are to the drive firmware, which now supports power management including a low power idle state.
The low capacities of the Optane Memory product forced Intel to position it as a drive specifically for caching in front of a much larger hard drive, but the Optane SSD 800p has enough space to serve as primary storage. While 64GB-class drives have disappeared from current flash-based SSD product lines, there are still plenty of 128GB-class drives around. These drive capacities certainly aren't roomy, but they are sufficient to install an operating system and several applications. For users that don't install huge AAA video games or deal with large collections of videos and photos, the 118GB 800p might not even feel too confining.
For flash-based SSDs, tiny capacities should often be avoided because they have much worse performance than larger models. The relatively small 128Gbit (16GB) capacity of a single 3D XPoint die means the Optane SSD avoids the limited parallelism that small flash-based drives suffer from, and the performance of a single 3D XPoint die is high enough that not much parallelism is needed to begin with.
Intel Optane SSD Specifications | ||||
Model | Optane SSD 800p | Optane Memory | ||
Capacity | 118 GB | 58 GB | 32 GB | 16 GB |
Form Factor | M.2 2280 B+M key | M.2 2280 B+M key | ||
Interface | PCIe 3.0 x2 | PCIe 3.0 x2 | ||
Protocol | NVMe 1.1 | NVMe 1.1 | ||
Controller | Intel | Intel | ||
Memory | 128Gb 20nm Intel 3D XPoint | 128Gb 20nm Intel 3D XPoint | ||
Sequential Read | 1450 MB/s | 1350 MB/s | 900 MB/s | |
Sequential Write | 640 MB/s | 290 MB/s | 145 MB/s | |
Random Read | 250k IOPS | 240k IOPS | 190k IOPS | |
Random Write | 140k IOPS | 65k IOPS | 35k IOPS | |
Read Latency | 6.75 µs | 7 µs | 8 µs | |
Write Latency | 18µs | 18µs | 30 µs | |
Active Power | 3.75 W | 3.5 W | 3.5 W | |
Idle Power | 8 mW | 8 mW | 1 W | 1 W |
Endurance | 365 TB | 365 TB | 182.5 TB | 182.5 TB |
Warranty | 5 years | 5 years | ||
Launch Date | March 2018 | April 2017 | ||
Launch MSRP | $199 | $129 | $77 | $44 |
The higher capacities that the Optane SSD 800p offers over the Optane Memory also allow for much higher write performance, which was the biggest weakness of Optane Memory. Still, this only brings the 800p up to performance levels slightly faster than SATA, with sequential write performance rated at 640 MB/s and 4kB random write at 140k IOPS. Read speeds are slightly faster than the 32GB Optane Memory and also look poor compared to flash-based SSDs, but Intel is specifying this performance at a queue depth of four, which is far lower than what most flash-based SSDs need to hit their peak throughput.
The addition of a low-power sleep state brings the idle power rating of the 800p down to just 8mW, compared to the 1W rating on the smaller Optane Memory modules. The endurance rating for both capacities is 200 GB/day for the five-year warranty period. Given the small capacity of the drives, this works out to 1.7 or 3.4 drive writes per day, which is considerably higher than normal for consumer SSDs.
The capacities of 58GB and 118GB look odd compared to the more usual amounts like 120GB or 128GB commonly seen for flash-based SSDs. The reason the 800p has slightly reduced capacity is that a 3D XPoint die's actual capacity really matches the nominal 128Gb, whereas NAND flash incorporates extra space above the nominal capacity to allow for error correction and wear leveling. For the Optane Memory, the difference between the power of two definition of 32GB and the traditional drive manufacturer's definition of 32GB provided sufficient space, but the 800p's metadata and error correction requires a bit more usable space be taken.
Pricing for the Intel Optane SSD 800p is similar on a $/GB basis to the Optane Memory, which is now significantly cheaper than the launch prices from last year. However, this still leaves the 800p as the most expensive consumer SSD on the market on both a capacity and per GB basis, with the 58GB model exceeding $2/GB. Even the ultra-high-end 900p is cheaper per GB than the 800p.
The Competition:
There aren't any close competitors to the Optane SSD 800p. Intel's Optane SSD 900p is a consumer-focused derivative of their enterprise Optane SSD DC P4800X and inherits its high power consumption and the large PCIe add-in card or U.2 form factors. The existing Optane Memory M.2 modules are closely related to the Optane SSD 800p, but their low capacities prevent them from being used for the same purposes.
Among flash-based SSDs, there are some current-generation 128GB-class NVMe SSDs but no 64GB-class drives. The small flash-based SSDs are all relatively low-end and far cheaper per GB than the Optane SSDs. The high-end NVMe SSDs that roughly match the 800p on price tend to have four times the capacity.
For this review, we are comparing the 800p against Intel's other Optane products and against a variety of flash-based NVMe SSDs ranging from entry-level drives to the premium Samsung 960 PRO.
Intel also sent us four of the 118GB model, so for the curious we have some benchmark results from using them in RAID. For those tests, the Optane 800p M.2 modules were installed in an ASRock Ultra Quad M.2 card and tested in our enterprise SSD test system, using Windows 10 and Intel's Virtual RAID on CPU (VROC) drivers. That enterprise test system includes all the latest firmware and OS patches for the Spectre and Meltdown vulnerabilities, so those test results reflect the overhead of those mitigations in addition to the overhead of the NVMe RAID software. The single-drive test results were all recorded on our usual consumer SSD test system that has not received any firmware or OS patches for the Spectre and Meltdown vulnerabilities.
AnandTech Enterprise SSD Test System | |
System Model | Intel Server R2208WFTZS |
CPU | 2x Intel Xeon Gold 6154 (18C, 3.0GHz) |
Motherboard | Intel S2600WFT |
Chipset | Intel C624 |
Memory | 192GB total, Micron DDR4-2666 16GB modules |
Software | Windows 10 x64, version 1709 |
AnandTech 2017/2018 Consumer SSD Testbed | |
CPU | Intel Xeon E3 1240 v5 |
Motherboard | ASRock Fatal1ty E3V5 Performance Gaming/OC |
Chipset | Intel C232 |
Memory | 4x 8GB G.SKILL Ripjaws DDR4-2400 CL15 |
Graphics | AMD Radeon HD 5450, 1920x1200@60Hz |
Software | Windows 10 x64, version 1709 |
Linux kernel version 4.14, fio version 3.1 |
- Thanks to Intel for the Xeon E3 1240 v5 CPU
- Thanks to ASRock for the E3V5 Performance Gaming/OC
- Thanks to G.SKILL for the Ripjaws DDR4-2400 RAM
- Thanks to Corsair for the RM750 power supply, Carbide 200R case, and Hydro H60 CPU cooler
- Thanks to Quarch for the XLC Programmable Power Module and accessories
- Thanks to StarTech for providing a RK2236BKF 22U rack cabinet.
Caveat: Many of our current SSD tests were not designed with tiny drives in mind. The results for the 32GB Optane Memory and the 58GB Optane SSD 800p do not represent exactly the same workload performed by the larger drives. Several of our synthetic benchmarks of sustained performance default to using a 64GB span of the drive, and in the case of the smaller drives, the test simply uses the entire drive. Likewise, the workloads represented by the ATSB Destroyer and Heavy tests don't actually fit on such small drives. The small drives still perform the same volume of reads and writes, but the block addresses in the I/O trace that are beyond the capacity of the drive are wrapped around to fit. The ATSB Destroyer and Heavy results for those two drives could be viewed as representative of the drive's performance as a cache device, but they do not include the effect of cache misses that would be present in a real tiered storage configuration.
116 Comments
View All Comments
Hurr Durr - Thursday, March 8, 2018 - link
Hypetane!iter - Thursday, March 8, 2018 - link
optane = hypetanex-point = xtra-pointless
It keeps getting worse and worse instead of getting better. The next x-point iteration may slip below nand even in the few strong points of the technology.
Also, it doesn't seem that enterprise is very interested in intel's offering, seeing how they struggle to cram the product in market niches where it is xtra-pointless, I'd go on a limb and assume that's not because of love for consumers or skipping on them fat enterprise product margins.
Also, it seems that intel gave very misleading information not only in terms of performance, but also regarding the origin of the technology. The official story is its development began in 2012 as a joint venture between intel and micron.
That however is not true, x-point can be traced back to a now erased from history company named Unity Semiconductors, which was flogging the tech back in 2009 under the CMOx moniker.
Courtesy of archive.org, there is still some trace of that, along with several PDFs explaining the operational principle of what intel has been highly secretive about:
https://web.archive.org/web/20120205085357/http://...
All in all, the secrecy might have to do with intel's inability to deliver on the highly ambitious expectations of the actual designers of the tech. It is nowhere near the 200% better than nand density, in fact it seems at the current manufacturing node it won't be possible to make more than 256 gb in m2 form factor, which is 8 times less than mlc nand or 24 times less than what was projected in 2009. Performance is not all that stellar too, a tad lower than what slc was capable at back in 2012, thank the gods nobody makes slc anymore, so there's a ray of sun to make xtra-pointless hypetane look good on paper.
chrnochime - Thursday, March 8, 2018 - link
Rambus renamed it to ReRAM according to this article in 2015, so it would seem the tech lived on through Rambus after the aquisition of Unity Semi.https://www.eetimes.com/document.asp?doc_id=132552...
But I'm not sure if it's the exact same tech as Intel's.
iter - Thursday, March 8, 2018 - link
Check the PDFs, what little intel posted about it is all there. They may have licensed the tech from rambus. It is not like rambus does anything other than patent milking anyway.iter - Thursday, March 8, 2018 - link
"Coincidentally", rambus bought unity in 2012, exactly when intel allegedly started developing...MDD1963 - Friday, March 23, 2018 - link
Not everyone remembers a few sticks of RAMBUS RDIMMS for some Pentium 3 boards costing $500-$600 a stick back in '99-'00....; and being outperformed by DDR. Nice job, RAMBUS!tommo1982 - Thursday, March 8, 2018 - link
Am I reading it right? Was Cross-point memory supposed to be cheaper than NAND?WinterCharm - Thursday, March 8, 2018 - link
Yes. But I guess we won't see that for a while.Latency and power consumption are great... but speed and capacity leave a lot to be desired. When MacBook Pros have NVME drives capable of 3.2 GB/s (yes gigabytes) at a 2TB capacity... Optane is far behind.
There are some advantages, but I expect that Intel will need to do a lot more work before these are cheaper, faster, and have higher capacity.
Reflex - Friday, March 9, 2018 - link
That said, latency is what users notice. Max speed is a rarely encountered scenario in most user workloads.iter - Saturday, March 10, 2018 - link
No human notices microseconds. Delay becomes noticeable at about 10-20 msec, depending on the individual's reflexes, becomes annoying at about 50 msecs, and becomes detrimental at 200+.10 mseconds is 10000 microseconds. Hypetane improves things in the double digit microseconds range. Humans cannot notice that, not today, not in a million years.