Asterox
Golden Member
- May 15, 2012
- 1,026
- 1,775
- 136
Yes, or another business loss for Intel.
Yes, or another business loss for Intel.
Yes, or another business loss for Intel.
An unfortunate thing indeed. The only issue with Optane was Intel's reluctance to officially support it on other platforms (like AMD platforms or ARM).
Intel could have done great things with the tech, instead they chose to kill it off. I guess they do still have quite a ways to fall before they realize what they are doing wrong.
The idiots calling the shots at Intel are simply too damn shortsighted. I really hope some consortium of investors (heck, even Chinese if that is what it takes) steps in and buys everything pertaining to Optane tech from Intel at a price they can't refuse (they need money badly anyway). Data centers are the primary target for this tech and if it is successful enough there, maybe it can trickle down into consumer stuff down the line someday.
After Optane is dead and gone, I am sure someone will test direct storage on a P5800X and the results will be amazing. The mining crazy showed Intel that gamers were willing to spend crazy $ on their gaming hardware, its too bad that they never found a way to get a gamer focused Optane drive into the market.
One of the big killers for Optane was also the price of doing the technology correctly. Intel's hybrid drives were a terrible implementation. Instead of making a hybrid drive that was essentially an Intel only solution with very limited compatibility, they needed to create a high end solution with universal compatibility. Imagine a new controller that managed tiered storage between DRAM cache, Optane cache and NAND. They could have created a drive for consumers that had similar performance to P5800X but fit in a M.2 port and it would have worked on any platform with NVMe. The secret sauce was always a product that combined Optane's insane 4KQ1T1 and NAND's insane sequential on one device that would just work on anything. I am 100% sure this was doable but the price, these drives would have been crazy expensive as we paid for both the Optane chips and the R&D on a new controller.
It is a dead end technology anyway (talking DCPMM). CXL would have killed it off eventually.
They never got the price down to a reasonable level. The only place it ever made some sense is in insane memory sizes (16-24TB sizes) which are super niche.
As a stand alone product for consumer, it had all kids of problem. Price/GB, capacity, form factor and heat were all issues compared to top of the line NAND solutions.
The only thing that came close was their hybrid drives but the way they work was offloading the actual caching computation onto the CPU and required Intel hardware and drivers to work. I'm sure everyone reading this understands the support nightmare that resulted from a failure condition. When Optane crashes, it can be a massive pain getting your data back and setting the cache back up again.
Makes you wonder, was there actually a plan that would have worked? Maybe let it cook for a few more years before the first commercial products?
They did the same thing to their own staff once before. There was a video up on youtube of an Intel summit where they specifically mention the 815P/M15 and a few online retailers had order forms up with Intel's official product identification numbers. The cancellation was already in the works when all of this happened.Huh, i really wonder if they were happy surprised.
- Oct 14, 2003
- 8,686
- 3,785
- 136
@nosirrahx Once Pat Gelsinger made the comment in an interview "I never want to be in memory", I had a bad feeling.
Remember Gelsinger had influences from Andy Grove, the famous Intel CEO and co-founder that changed Intel from a memory focused manufacturer to a CPU based one which paid off handsomely for them.
I think the best way would have been to make Optane Memory really like memory and use the server Memory Drive feature to extend your system memory. I think that would have had lot more appeal. Then that would eventually transition into PMM modules.
I think another thing that killed it was the endless Sapphire Rapids delay. I seriously don't get it. Open it up so it can work on AMD servers! That's how you expand into a new market, not let it's fate be tied to a problematic platform.
Yea not opening up was an issue but Intel themselves had serious issues making it work on their own platform.
Remember Gelsinger had influences from Andy Grove, the famous Intel CEO and co-founder that changed Intel from a memory focused manufacturer to a CPU based one which paid off handsomely for them.
I think the best way would have been to make Optane Memory really like memory and use the server Memory Drive feature to extend your system memory. I think that would have had lot more appeal. Then that would eventually transition into PMM modules.
I think another thing that killed it was the endless Sapphire Rapids delay. I seriously don't get it. Open it up so it can work on AMD servers! That's how you expand into a new market, not let it's fate be tied to a problematic platform.
Yea not opening up was an issue but Intel themselves had serious issues making it work on their own platform.
Optane was supposed to herald an age of computing where computers would be almost instant-on without any perceptible boot delay. Guess that's going to remain the stuff of science fiction for the foreseeable future.
TRENDING THREADS
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 10K
-
Discussion Speculation: Zen 4 (EPYC 4 "Genoa", Ryzen 7000, etc.)- Started by Vattila
- Replies: 13K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 7K
-
Facebook
Twitter