Sunday, May 29th 2022
Phison Showcases 12 GB/s Speeds for PCIe 5.0 SSDs Through Its New E26 Controller
Phison has showcased the expected performance of its upcoming PS5026-E26 controller, built to usher NVMe SSDs into the PCIe 5.0 realm. The company showcased its new controller's prowess by building a reference SSD design based on 1 TB of Micron's TLC NAND. Phison's new controller has been built from the ground-up to accelerate next-generation SSD workloads - including direct access technologies based on Microsoft's DirectStorage API, accelerated by two ARM Cortex-R5 cores and three proprietary CoXProcessor 2.0 accelerators built on TSMC's 12 nm process.
Phison's internal testing shows its reference SSD achieving sequential read speeds of over 12 GB/s in CrystalDiskMark, with sequential writes going as high as 10 GB/s - a 70% performance increase compared to the world's fastest PCIe 4.0 SSDs, which currently top out at around 7 GB/s sequential speeds. As to 4K performance, one of the most tangible metrics for user experience, random reads are set at around 16.000 IOPS, showcasing room for improvement with further firmware optimizations for actual shipping products.
Interestingly, Phison opted for the M.2 2580 form-factor for its proof-of-concept SSD, which features a slightly wider PCB and connector footprint that's not backwards compatible with M.2 2280 slots. Expect SSDs based on Phison's PS5026-E26 controller to hit the market later this year - closer to AMD's release of its 600-series chipsets for its next-generation AM5 platform.
Source:
TechSpot
Phison's internal testing shows its reference SSD achieving sequential read speeds of over 12 GB/s in CrystalDiskMark, with sequential writes going as high as 10 GB/s - a 70% performance increase compared to the world's fastest PCIe 4.0 SSDs, which currently top out at around 7 GB/s sequential speeds. As to 4K performance, one of the most tangible metrics for user experience, random reads are set at around 16.000 IOPS, showcasing room for improvement with further firmware optimizations for actual shipping products.
60 Comments on Phison Showcases 12 GB/s Speeds for PCIe 5.0 SSDs Through Its New E26 Controller
These PCIe 5 drives are going to be a bit pointless really imo. A fast PCIe gen 4 drive is plenty fast enough imo. It's the same as when ADL came out and everyone pooed at Intel for putting a PCIe 5 GPU slot now AAMD are doing the same without there even been any video cards that support it. I suppose you can use it for expensive first gen PCIe 5 SSDs though at least.
Getting your legs stuck in pants is just... so special.
(And no it's not an obesity thing, just that for 5 years or so i was riding 15KM to work, working in tobacco fields and vineyards, and then riding home)
So the real reason RAND4K is slow is because phison cant find pats that fit
Personally I'd be infinitely more excited about high capacity cheap SSDs. A manufacturer of a reasonably priced 10TB SSD would absolutely get my money - make it 3.5" and SATA to cut costs, I don't care. I just want to get rid of the unreliable and noisy spinning rust nonsense from my life.
The most impressive thing from this showcase to me is they got a toolless install for the m.2.As we see with 500hz monitors, marketing rules the roost, as you said bigger numbers.
EDIT: If the connector is indeed wider, it should at least be backwards compatible (Gen 5 slot, Gen 4 SSD) ...
Also, is space on (some or many) motherboards, or PCIe adapter cards, so constrained that a 3 mm wider PCB wouldn't fit in its place?
So on a bare PCB there should be no restrictions mounting 25mm to a 22mm slot.
However on boards where 22mm wide spacing is assumed by motherboard covers there may be compatibility problems going forward.
As much as this tech annoys me by changing the dimensions, at least its fully backwards compatible and not REQUIRED to be wider - we might get lucky with fat 5.0's in 4.0 devices, if theres enough clearance
If the electrical connector is changed and it stops being compatible, i'll have a tantrum
Why are they even so low? is it something to do with the interface or OS or what?
SSDs are more complex beasts, their controllers do parity and error-checking, encryption, compression/decompression, and now conversion from SLC to TLC/QLC all simultaneously. Even the highest-grade NAND is riddled with manufacturing defects so the controllers have to work around these. NAND is fragile with limited program/erase cycles and so minimising the amount and number of times written to it is vital. All of this NAND management to compensate for flaws in the technology takes time so a good SSD controller is 50x slower than Optane, with the best controllers on the market barely able to break 500microseconds in a best-case scenario.
If they wanted to, companies like Samsung could make NAND controllers that were lightning fast like Optane but they'd wear out the NAND in a couple of weeks and the data would be constantly corrupting - you'd be lucky to get back what you put in most of the time. So realistically, the NAND management required to make NAND a viable storage technology is enormously complex, and cannot be omitted. Until flawless yields of 100,000 P/E-cycle NAND that have near-zero defects are common, we're going to have this 0.5-1.0ms latency on SSDs that caps IOPS and 4K performance.
I guess we will be stuck with slow RND4K for a long time by the looks of it.
EDIT
Have to say the load times on Sniper Elite 5 are the quickest i have seen for a long time, the game is on one of my WD SN850's
It's not quite RAM, but it shares much more in common with RAM than it does with NAND, including the cost per gigabyte :laugh: