Intel Optane SSD 900P Review: 3D XPoint Unleashed

  • Thread starter Thread starter Guest
  • Start date Start date
Status
Not open for further replies.
Can someone comment on the statement "your shiny new operating system was designed to run on an old hard disk drive. "? I somewhat understand it, but what does that mean in this case? If microsoft wanted... could they rewrite windows to perform better on an SSD or optane? Obviously they wont until there are literally no HDD's on the market anymore due to compatibility issues but I just wasnt sure what changes could be made to improve performance for these faster drives.
 
Thanks for the review. I have a couple questions if you don't mind answering them.

1. Is PCI-e 3.0 a bottleneck for Optane drives (or even flash in general). If solid state drive developers built PCI-e 4.0 drives today, would they scale to 2x the performance of modern 3.0 drives assuming there were compatible motherboards?

2. Can someone explain what queue depth is and under what circumstances it's most important? The benchmarks show WILD differences in performance at various queue depths. Does it matter that flash based drives catch up at greater queue depths? Is QD1 the most important measurement to desktop users?

3. In the service time benchmarks, it seems like there is no difference between drives in World of Warcraft, Battlefield 3, Adobe Photoshop, Indesign, After Effects. Sequential and Random Read and Write conclude "Optane is WAAAAY better" but the application service time concludes "there's no difference". Is it even worth investing in an Intel Optane drive if you won't see a difference in real world performance?
 


Simplest way to answer is the term, "lowest common denominator". They(MS) designed 10 to be as smooth and fast as a standard old platter HD could run it, not a RAID array(which a SSD is basically a NAND raid array), or even better an XPoint raid array. Short queue depths(think of data movement as people standing in line) were what Windows is designed for, cause platter drives can't do more data movements than the number of heads on all the platters combined without the queue(people in line) growing, and thus slowing things down. In short, old platter drives can slowly handle like 8-20 lines of people, before they start to clog up, while nand, can quickly handle many lines of people(how many lines depends on the controller and the number of nand packages). Admittedly, it may not be the best answer, but it's how I visualize it in my head.
 


That helps a lot actually! thank you!
 
Queue depth can be simplified to how many operations can be started but not completed at the same time. It tends to stay low for consumer applications, and get higher if you are running many virtualized servers or heavy database access.

With spinning metal disks, the main advantage was that the drive could re-order the queued requests to reduce total seek and latency at higher queue depths. With NMMe the possible queue size increased many-fold (weasel words for I don't know how much), and I have no idea why they provide a benefit for actually random-access memory.
 


Off the top of my head, the entire caching structure and methodology. Right now, the number one job of any cache is to avoid accessing the hard drive during computation. As soon as that happens, your billions of cycles per second CPU is stuck waiting behind your hundredths of a second hard drive access and multiple second transfer speed. This penalty is tens of orders of magnitude greater than anything else, branch misprediction, in-order stalls, etc. So anytime you have a choice between coding your cache for greater speed (like filling it completely to speed up just one program) vs avoiding a cache miss (reserving space for other programs that might be accessed), you have to weigh it against that enormous penalty.
 
where is the power usage test ? I can see a HUGE heatsink on that monster , and I want the Wattage of this card compared to other PLEASE.

This SSD cant be made M2 card . so Samsung 960 pro has a huge advantage over it.
 
A lot of the charts that show these massive performance gains are captioned that this is theoretical bandwidth which is not seen in actual usage due to the limitations of the OS. I get it, it's very cool, but if the OS is the bottleneck, no one in the consumer space would see a benefit, especially since NVMe RAID is still not fully developed. If I am wrong, please correct me.
 
Chris - I clicked on the Tom's Hardware link to the newegg page. The specifications are showing PCIe 4.0. I know the new PCIe 4.0 standard has been approved and that we will see PCIe 4.0 products this time next year. Any chance the 900P is PCIe 4.0 ready or is it just a mistake?
 


We were told 24 hours before the review that it is PCIe 3.0 X4. I'm not sure where the 4.0 info comes from but more than one seller has it listed.
 
I have to admit I get confused by what all the different tests would mean to my day to day experience. The biggest issue for me is saving large 3D and Photoshop files. Some of my files are over 1GB and take a long time to save. Since software crashes occur I need to save frequently and each time I do this it creates a serious interruption in the workflow. Which of the tests best represents the speed of saving a single, massive file? I've heard many times people say that upgrading from an HDD to and SSD didn't really change the speed of saving files. Is this type of save bottlenecked by the CPU or something else? If someone could shed some light on this it would be much appreciated.
 


That's not entirely true. People will benefit. It's still faster than NAND SSDs in everyday scenarios, just the amount you notice big increases are decreasing.

First, NVMe RAID sucks, especially for Optane. RAID adds software-induced latency and the point of these drives are low latency. You reduce some of the gains by RAID-ing it.

Optane SSD
-No need for TRIM. TRIM is a requirement for non-Optane SSDs because the media is slow and TRIM, extremely roughly speaking is like defrag for SSDs
-SSDs slow down drastically when drive is full. Not with Optane.
-SSDs slow down when it's "dirty". That's every time when the drive is loaded and has no time to TRIM, or the demand is high that controller and the buffer gets overloaded. Not with Optane
-If you just erased large amount of files, you'll see stuttering with SSDs. Not with Optane

Not to mention when you are transferring or working with numerous small files, the speed advantages of Optane will be enormous.

With games and applications the benefits will vary because some applications are CPU-bound, and some portions of code can't be accelerated much because other parts of code have fixed loading times. For example, you can't make ads disappear faster by having a faster drive.

Whether it's worth the price of Optane? That depends on you. I think for most users its not. For users that want to just get the fastest system, not having an Optane SSD will be kinda strange. I mean, I expected $1000+ for 480GB. At $380 plenty of enthusiast systems can get the 280GB version.
 
It's obvious that once an SSD is installed in a system, the game loading times become CPU and/or RAM bottlenecked, not storage bottlenecked. That's why these fancy NVMe SSDs barely load games faster than budget SATA SSDs. Can Tom's Hardware please start testing how different CPU platforms/architectures affect game load times? I bet an i7 8700k with a budget SATA SSD will load World of Warcraft much faster than an FX 8350 with this Intel 900p.
 


Yeah, basically the operating system are limited to fast operative memory and slow storage for historical reasons. Now, the storage is fast enough to make feasible its usage in more RAM-like manner. A good example would be the file mapping features available in every modern OS. With slow storage this needs copying to RAM of chunks of the file and making that piece available at certain address. With much faster storage you can skip the RAM entirely. If you start thinking in that direction it seems quite interesting to develop a 3-d kind of storage support in the operating systems. Like the file mapping there are a number of existing OS features that can benefit directly and become the actual parts of the OS that are clients to this 3-d kind of storage.... Well, I mean that when you think of it, we already have enough constructs in the modern operating systems to handle this and we can expect more efficient methods rather sooner than later - no major redesign will be required to feel the benefits.
 
Since Optane doesn't require TRIM or garbage collection it would be cool to see some non-AHCI SATA 1 drives or maybe even PATA/Compactflash cards that use it. It would be great for older computers that don't have TRIM and want solid state performance without degradation. I know it will never happen, but I can dream.

Edit: Or even just a SATA 3 version that I can put in my PS4, since PS4 Pro has SATA 3, but doesn't support TRIM.
 


As you stated that wont happen, going to SATA would incur a big performance hit and increase latency hugely compared to NVMe. Just get a SSD for yoru PS4 that has internal garbage collection capability's (most do) so you don't have to rely on TRIM.
 

That internal garbage collection doesn't work. The data has to be overwritten with zeroes for the internal garbage collection to think the data has been cleared. Most OS's just delete the pointer to the file in the file system, but don't actually overwrite the data, so even if I delete all the games off of my PS4's SSD, the SSD still thinks it's full and can't do garbage collection. And yes the SATA interface would have higher latency, but maybe high latency lower binned chips could be used or something.

 
Question is, does this SSD work with a ASUS z170 pro-gaming Mobo. Asus and Intel support could not confirm it.... I am worried.
 


So what? Look at the sequential results. Optane with "just" 2500MB/s reads beat the 960 with 3200MB/s reads in the most important QD1-4 depth.

Vast, vast majority of consumer usage, including games are on QD1-4. That's just sequential. When you do small file transfer it will kick any NAND SSDs, including the 960 EVO to a pulp.
 
Status
Not open for further replies.