What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

Intel Optane 900p

AkG

Well-known member
Joined
Oct 24, 2007
Messages
5,270
Isnt Linus the one who recommended SATA to 6-pin adapters for PCIe risers to MINERS?!.. ie he recomended a setup that can burn your house down. He should stick to what he is good at... opening boxes.

The 900P is a killer drive. So good even the small 280GB is basically as fast as the larger 560GB'er. THAT is how good the 900P is. If you can afford it and can live with the small capacity its a no-brainer.

Edit: ROFL. He couldn't even get the name right. Check out the name he gives in the comments section: '900p'. He doesn't even understand Intel's latest nomenclature... as small p means NAND, large P means Optane. Sooo he is talking about Optane but is telling everyone its NAND. If he can't such obvious details right... why would anyone expect him to get the nitty gritty right? :shok:
 
Last edited:

belgolas

Well-known member
Joined
May 9, 2007
Messages
4,359
Location
St. Thomas, Ontario
While he is certainly not the best reviewer he at least does comparisons for what I care about. Every Intel Optane 900P review I have looked at don't show us much of real word benchmarks. Intel marketing tried to sell this to enthusiast gamers so I want to know if this make any difference to gamers. Turns out it doesn't change game load times or make the computer feel any faster. So while this is the fastest ssd it isn't worth it in my opinion because for the cost you can get a larger ssd.

This is the problem I have with a lot of reviewers. They show off the synthetic performance and some benchmarks but I am not a professional that would use this for some job but just for my hobbies which is mostly gaming. I get excited to read a review for example on new DDR4 at extreme frequencies like 4400 mhz only to find that the reviewer doesn't show any gaming performance. Sure it is neat to know that "x" product is faster but I don't care if it doesn't improve my experience.
 

AkG

Well-known member
Joined
Oct 24, 2007
Messages
5,270
Gotcha. Its all good. I just dislike seeing sloppy work for the sake of click whoring... especially when he gets IMPORTANT details (like not burning your house down) when making recommendations to novices. Its an on going pet peeve of mine as he has been pulling that shit for years and years now. :/

Any modern x4 NVMe PCIE drive is going to be bloody fast for gaming. Its really not the market for 9-series SSDs. It will be faster... but at a certain point peeps have to ask themselves 'how fast is fast enough?'

NOW with that said. Remove the CPU and RAM bottlenecks and it will be faster... as the 9-series is so bloody fast 6 and even 8 core CPUs are the bottleneck. Stick in in a 10 to 14 core beast with 64GB of fast ram and watch it really spread its wings and fly..... So HEDT or nothing is how I look at uber $$$ drives these days ;)
 

Marzipan

Well-known member
Joined
Nov 21, 2007
Messages
12,145
Location
Prince Rupert, British Columbia, Canuckistan
Intel SSD cost to much and don't perform as well as others. and most SSD's don't suffer much in performance differences depending on capacity anymore, not like they used to.

as much as I dislike Samsung...their 960 Evo or Pro is king. :whistle:
 

EmptyMellon

Well-known member
Joined
Nov 6, 2010
Messages
524
The one thing I have to pick on in that 900P LTT review video is SolidWorks' benchmark numbers (4:40 mark). It is clear he is not using a workstation/certified GPU (no "RealView" value) and as such, it does not reflect well on the performance of any NVMe storage type. 30 to 34s is slow as snails! [I had better I/O numbers in SolidWorks 2017 (SP 3.0) using a Samsung 840 Evo under an i7-3820 CPU and a K4000.]

Just for comparison here are my current system's SolidWorks 2017 (SP 5.0) benchmark numbers (same system, different workstation GPUs) [lower is better]:

Graphics :NVIDIA Quadro K4000 (PCIe 2.0)
SolidWorks Preformance Test (sec) (overall= cpu + Graphics + I/O)
Overall=66.9 Cpu= 35.6 Graphics=8.0 I/O=23.3 Render=0 RealView=6.7 Simulation=0.0

Graphics :AMD Radeon Pro WX7100 (PCIe 3.0)
SolidWorks Preformance Test (sec) (overall= cpu + Graphics + I/O)
Overall=22.8 Cpu= 0.5 Graphics=5.1 I/O=17.2 Render=0 RealView=4.3 Simulation=0.0

So even a Quadro K4000 with a PCIe 2.0 bus interface has a better I/O than a GAMING 1111 RGB FAST-FAST GPU (most likely a 1080 Ti). Use a proper certified (workstation) GPU and then show the numbers. The I/O device is bottle-necked by the poor GPU. Perhaps Intel should have a word with LTT about that. I would love to see those numbers, when done right.

Intel SSD cost to much and don't perform as well as others. and most SSD's don't suffer much in performance differences depending on capacity anymore, not like they used to.

as much as I dislike Samsung...their 960 Evo or Pro is king. :whistle:
It is not king with their current 3B6Q firmware debacle. There is a lot of unhappy people out there at the moment. Hopefully they won't screw it up again with their coming fix.
 

Oldnconfuzed

New member
Joined
Mar 1, 2018
Messages
2
From the perspective of the customer

belgolas has it right.

As a many-year fan of Hardware Canucks, I hope what follows will be taken as constructive criticism.
I’m a fairly affluent gamer; and I too have read perhaps 8 reviews of Intel’s new Optane 900P.

As noted, Linus’ “The fastest SSD for gaming, and one big problem” is indeed an attempt to meet my needs. (The 30 seconds of video; two game review starting here)

The key problem is that any reviewer looking at a “new piece of hardware” spends an inordinate amount of time looking at ways to
• comment on architecture/design;
• benchmark it against others;
• stress test it; and
• provide “real world performance” comments.
Let’s call this the “hardware” focus.

Problem is, I’m a customer … one with cash … and want to know what value I’ll get from spending that cash. A review that explains this concisely and clearly would be “customer” focused.

There may be more, but from my personal experiences there are three types of customers for hardware like the Optane 900P:
  • Gamers: people who want seamless gaming performance; people (high end) who would pay through the nose for an advantage.
  • Video Production /Streamers: a likely smaller group, who have a specific need for fast and cost efficient video production. An even smaller subset who Game & Stream at the same time.
  • Data Hogs: the “big data” folks. Database manipulation. Copying impressively large amounts of data from one place to another. Perhaps the “encrypters”; perhaps the “data compression” folks.

Obviously I’m outside my comfort level with the 2nd and 3rd of these … so let’s just re-focus on gaming.

What value can an SSD provide to a gamer?

No compromise gaming performance (with multiple internet web-pages open; probably a spreadsheet or two as well). What does THAT look like?

  1. Faster “start-up”. I want to press “go” on my computer and get into my game in the minimum time possible. Load windows; load my game; download updates to my game.
  2. Reduced load screen times. This is HUGE for me. I may, in reality, spend only a small fraction of my gaming experience in loading screens, but it SEEMS like 25% or more because staring at the equivalent of a spinning hourglass adds zero value to the gaming experience! I need to change from one character to another, then zone from one location to another. A minute or more staring at load screens (with 2 older SSD’s in raid 0). THAT’s what makes my gaming experience suck. THAT’s what I would pay money to fix.
  3. Capacity: The page 3 notion that I’d lose nearly 20GB of a 280GB drive is indeed relevant to assessing whether the drive would meet my needs (although I didn’t quite follow; nor did I see reference to the larger 480GB drive’s stats).
  4. Consistency: I expect to know before I buy “hardware x” whether it will throttle or not.
  5. Reliability: The page 2 “durability” comments are valid and meaningful. Appreciated. My current raid 0 SSD configuration scares me … a lot!
  6. A “dollar value proposition” to go along with those metrics. Would I spend $300 more for a 0.5% improvement? Maybe; maybe not.

There’s not much more I want (SSD wise). Expand the list a bit to “value propositions” presumably “not applicable” to SSD’s:
  • The pipeline “from game server to PC”: IPS/Modem/Router/(Wifi?): The best possible internet experience with Low latency/ping; low dropped packets. How many router reviews quote comparisons of ping?
  • Butter smooth video (call it FPS+; as it would include latency and anything to improve the visual experience such as accurate colour reproduction) How many screen reviewers rank displays by latency?

... all with "test scenarios" that include a few open web-pages; a separate communication system (like Teamspeak or Discord) and a couple fairly large spreadsheets would be helpful as well, since just testing a single game in Windows isn’t reflective of my real-world workload.

So back to the review in question for the Optane 900p:

First, the competition. The Samsung 960 Pro is reputedly one of, if not the fastest SSD’s available … and price-wise the most viable “competition of note” for the 900p … yet despite references to it on page 2 of the review (and verbiage on page 3 directly comparing sequential write performance), it’s not in the comparison tables. Does that help me with my purchase decision? Same for the Samsung 850 mentioned on page 1. Does not help one bit. My money would NEVER go to any of the competition in the tables … making them largely or wholly irrelevant. Worse still, most are SATA drives; a fundamentally different technology from a generation ago that influences the result of the table comparisons MORE THAN the hardware themselves).

Second, the Consistency argument (3 above). Reviewer was “shocked” at the “significant misstep” of potentially inadequate cooling. Perfect valid concern. Any empirical assessment of throttling? Nope. Do I leave the review with an understanding of short or long term performance sustainability in a gaming environment? Nope. Any point to citing cooling concerns if they have no real-world applicability? None that I can see. At the very least it would have been of value to me to know that for certain though.

Third, the value proposition relative to gaming needs. This is the one that truly irritates. I don’t run ATTO; CrystalMark; PCMark8; etc etc. and I most certainly don’t WANT to run them. They add zero value to my life.

I can understand the notion that “if it’s the best there is, use it”. It takes time and money to conduct these tests, yet in NO WAY do I understand at the end whether I should be buying this piece of hardware or not. 120 web-pages open in Firefox? Let’s see … the last time I tried that was … omigod … never!! (has anyone, ever?) If you instead told me it would reduce windows load times from 20sec to 15sec? I’d “get” that. If you told me it would reduce my game’s initial load and/or zone change screen loads from 40sec to 30sec? I’d “buy” that (literally; and today!!). I’d also never miss a review by that reviewer!


So I also get that this would be hard to do as a reviewer. Which games to choose? Games change each year, making benchmarking difficult. Other hardware and other factors can influence results … making true comparisons to “my system” challenging. Yet I still don’t care. ANYTHING that gave me answers to the “value proposition” of spending my cash on a piece of hardware is preferable to looking at stats that don’t tell me if my life would be better or not.

I AM the esport gamer at the top of page 2’s pyramid! I certainly can afford to buy one … but won’t waste money on something that adds virtually no value.

As an aside, I’ve also spent a fair amount of time trying to figure out what bottleneck(s) are causing excessive screen load times … without much success. THIS (and ways to mitigate it) would be of enormous interest.

I also know all too well that like my moniker, I’m old and senile. I could be dead wrong on specific points; or wildly missing other valid ones. Yet I still believe the premise is valid:

Please review hardware from the perspective of the customer, not the hardware.
 

AkG

Well-known member
Joined
Oct 24, 2007
Messages
5,270
Thanks for the input mate. Just a few points of clarification. We do exactly that. That is why there is different awards.

Dam Good = best of the best usually without concerns over budget. EG Intel 900P
Dam Innovative = Innovation that may or may not make the new tech the best NOW but has moved the entire industry forward. EG Crucial MX300 and their CuA (and to lesser extend MX500 and second gen CuA NAND)
Dam Good Value = Its a great value for customers. EG Intel 760p.

On screen load times. It could be the storage, the CPU or the GPU... could even be lack of RAM. What is your rigs specs? Might be able to help narrow down the bottleneck. Just understand some 'screen load times' are artificial... as in the game designers added them in for 'cinematic' purposes... not because your system is slow. Worst offenders/ best examples of this is CoD games.

On thermal limiting. It is too variable to show that if in X system it will happen X times. Rather the results themselves of ALL tests show its impact. On OUR testbed, on OUR sample. Will happen in a YOUR case with YOUR sample. Maybe, maybe not. Depends on how hard you stress it. The heat sink is more of a heat spreader than a true heatsink as its cooling fins are mostly covered. Stick a good fan near it and it wont thermally limit... (nearly) ever.

On real world examples. I do that all the time in FireFox. I have anywhere from 200 to 400 tabs open in FF, another 2-4 hundred in Waterfox... and when really doing research another 1-300 in chrome open at any time. More to the point. Using fewer pages would simply LOWER all the times and make it harder to tell one drive from another. IE my test setup spreads the drive results out more to reduce margin of errors. Think of an old school speed trap that uses a stop watch and two set points. Move the points further apart == more accurate results. Closer together... and you may get a ticket for doing the speed limit. :thumb:

On synthetic tests. They show how much potential performance a drive has in an easily reproducible manner. The real world shows how much of that potential is actually useful to buyers. Without one or the other an accurate assessment can not be made.

On Samsung. They wont send me samples. As this is a hobby I refuse to pay for free advertising for mfg'ers. Thus they are not in my charts. Yes I have discussed this issue many, many times over the years with the owners of HWC. Nothing I can do about it. I am mfg'er agnostic and always give an honest opinion on any sample sent my way. I never do 'hatchet' or 'cheerleader' reviews. Some manufactures don't like taking that risk. They want only positive results. If you feel strongly about a drive missing from the charts. Drop the owners a line. Easiest way to do that is via the Editor In Chief - Skymtl.

Hope that helps.
 
Last edited:

Oldnconfuzed

New member
Joined
Mar 1, 2018
Messages
2
Fair enough

Appreciate the reply to my rant (a lot). Thank you. I really had intended it as a general rebuke to reviewers; not to you personally ... although on reflection, I appreciate some of the comments were indeed too direct.

On Awards: I didn't consider the "lack of an award" as a data point. Fair enough though.

On Screen Load Times and my rig specs: That was helpful, thanks. I incorrectly assumed the culprit was storage alone. I run two (older; non-high-end) SSD's in Raid 0 ... and everything I read has led me to believe the potential practical improvements from newer tech (be it NVMe M.2; 3D XPoint; or what's around the corner) to be minimal in real world testing. As to my rig, it would be unhelpful as I'm looking to replace it, entirely (hence my intense interest). The good news? If indeed it is a mix of everything (CPU; GPU; Storage; RAM), there may well be a happy ending for me. I have severe deficiencies in both CPU and GPU right now. Fingers crossed. (and super appreciate the offer to advise).

On 800 tabs open, two words: holy crap! :shok: (well, two words and an exclamation point). Other applicable two word replies? "but how?" ... and "deeply disturbing" :haha: ... yet I'll stick with "holy crap!" Before I fold completely on this point though ... please consider my input on "types of customers". I still have to contend that "massive users of more tabs than my current computer could dream of opening" is a rather small market for SSD's ... but clearly this is only my opinion. Way To Fold

On real world testing: Hope it's ok that on this one, we can agree to disagree. :canadianwave: (kinda liking the crazy icons; #CanadianApology)

On Samsung & not being able to get the "debatably right" comparative equipment: Yikes; I didn't even think of that. I certainly do get that "keeping/having one of everything" available is problematic and cost prohibitive. And yes, I certainly avoid reviews and reviewers who do the "cheerleading thing" to keep in the manufacturer's good books ... so kudos to you and keep it up. I think we can agree though: it just plain sucks.

Thanks again for taking the time. I learned a bunch.
 

AkG

Well-known member
Joined
Oct 24, 2007
Messages
5,270
Its all good mate. All input is good input. I don't things personally. So it is all good. :)

Yes I dig deep when doing research... and go down some odd rabbit holes until I understand exactly what the pros and cons of a new tech are. Takes awhile but worth it. Large data analysis is what I was trained to do 'in the real world'... so its old hat to me to have 100s of POVs on a subject active at anyone time.

Its funny but the FireFox addition was not my idea. It was a request from multiple long term readers a few years back. I took their input and simply ran with it. So if there is something missing from the reviews just let me know. No guarantees on adding it in, but I will think about it and see if its practical. All suggestions are welcome!

When RAID'ing SATA SSDs on older rigs your PCH is going to be a bottleneck. Plus you are maxing out at about 1GB/s. Which is 50% to 300% slower than what NVMe can do (1.5 to 3GB/s). Will it make a 'huge' difference with this extra speed? Probably not. Your mobo and PCIe interface, the slow CPU, the slow RAM... all could be slowing things down and be the 'big' bottleneck. So I would not recommend sticking a modern NVMe drive into an older rig.
 

Latest posts

Top