One big use may well be searching - telling a SSD to find all documents containing a specified string for example. This is a job that is CPU and I/O intensive on a traditional system.
That's so true. I'm still shocked every time I put a search query into explorer and windows takes ten minutes to spit out the results. Feels like it hasn't changed in at least 10 years.
One more I saw hardware implementation ,that will save Windows ASS ,whit very obsolete NTFS,MacOs searches are insanely fast whit minimal overhead for the computer.
I'm a noob so please correct me if I'm wrong: it sounds like this is a product intended for server with typically huge number of drives or SSDs but limited CPU cores, no? Isn't the windows example limitation from the OS? Most desktop users will probably have more cores than parititons; sorting out a search from couple/few partitions sounds like a software limitation?
That is a $7000 FPGA which is 3/4 filled with the logic to drive the PCIe and SSD interfaces, which makes the discussion of 'servers with dozens of them in' quite alarmingly expensive; also leads me to wonder why they've integrated it with $800 worth of flash. I am also quite surprised not to see a QSFP+ port on the SSD. I'm sure there is some really exciting use case that they will write a white paper about, but this is really very esoteric for the front page of anandtech.
BittWare make an accelerator called the 250-U2 which seems to retail for ~2800$ with 8GB of DDR4. They also have a PCI-E card version with 3.84TB of storage which goes for ~6600$. Samsung cuts out the overhead by selling their own Flash, so I wouldn't be surprised if MSRP was closer to 5-6k$.
I'm not at all surprised there's no QSFP+. U.2 is a great form-factor for both compatibility with existing storage server designs and handling the thermals.
That said, people seem to be missing the point of these devices. If you operate on large datasets, being able to optimize access at the device level can significantly reduce the amount of computing power needed to do the same operations with a CPU in RAM, especially when you are using an FPGA to implement parts of the hardware. The cost of one of these easily outweighs the cost of having multiple NVMe drives to satisfy the same I/O requirements and then having to rely on sub-optimal CPU algorithms for processing that data. This will be a huge boon to database and search engine performance. You could also use them to implement rather large content-addressable memories for networking applications and the like.
"That is a $7000 FPGA which is 3/4 filled with the logic to drive the PCIe and SSD interfaces"
The SSD interface doesn't run on the FPGA - just the PCIe switch. There's no indication that it's "3/4" full, and in fact that seems like a bizarre assumption to make given the obvious downsides that you note.
Thinking this would be a great place for bored systems administrators to hide low end gaming hardware on corporate servers. One more thing we have to figure out how to monitor and secure also so its certainly a risk-vs-reward critter.
There's a "How to buy" link, but it presumes that you know "how to use" (you would want to have experience programming FPGAs). 4TB seems to be a rather tiny offering, but there are already Cloud Services that you can test out these drives.
There are a lot of use cases listed on the above links, but if you can't justify U$7K for a very smart 4TB of storage this wasn't made for you; try the Cloud Services already setup.
So 6k euros for 1.1M LE + 4GB ram + 4TB storage. Quite expensive. The NiteFury I am using now was 300 euros for 200k LE + 512MB ram. Perhaps this is worth it, actually. Thank you for the mouser link!
> There are a lot of use cases listed on the above links, but if you can't justify U$7K for a very smart 4TB of storage this wasn't made for you; try the Cloud Services already setup.
I read this as someone is already renting a cloud instance with one of these installed. As you seem to conjure up links I could not find, is there a link for this service as well? Does Hetzner have these or ?
Nimbix has a page up about having SmartSSDs in their public cloud, but it doesn't appear they've added that option to their price calculator: https://www.nimbix.net/samsungsmartssd
Thank you - so not available, and the SSD logic takes 800k out of the 1.1M LEs, so this is basically not a real contender for any price above $1000 for me. Too bad, it seemed great at a first glance.
Samsung should use ASIC PCIe switch rather than using LUTs on the FPGA. FPGA is for programming, the PCIe switch is a given. But I guess Xilinx just want user to design some DSP only algorithm, which is not totally unreasonable.
There aren't a lot of options for small 12-lane PCIe switches. The PLX PEX8714 is 19x19mm and 2.7W, and ASM2812 is 21x21mm. I'm not sure how easy it would be to squeeze one of those into an already crowded U.2 drive.
It's cool tech for sure, but I am not sure I get why this has to be implemented as an FPGA, and not exactly a small and cheap one, either? Could this be done using, let's say, in software with a decently fast CPU/NPU combo? Why does it have to be implemented as field-programmable hardware?
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
25 Comments
Back to Article
SaberKOG91 - Friday, November 13, 2020 - link
Good to see this finally come to fruition on the hardware side. Here's hoping software support catches up.Duncan Macdonald - Friday, November 13, 2020 - link
One big use may well be searching - telling a SSD to find all documents containing a specified string for example. This is a job that is CPU and I/O intensive on a traditional system.vol.2 - Friday, November 13, 2020 - link
That's so true. I'm still shocked every time I put a search query into explorer and windows takes ten minutes to spit out the results. Feels like it hasn't changed in at least 10 years.ClaudioMP - Sunday, November 22, 2020 - link
One more I saw hardware implementation ,that will save Windows ASS ,whit very obsolete NTFS,MacOs searches are insanely fast whit minimal overhead for the computer.K_Space - Sunday, November 15, 2020 - link
I'm a noob so please correct me if I'm wrong: it sounds like this is a product intended for server with typically huge number of drives or SSDs but limited CPU cores, no?Isn't the windows example limitation from the OS? Most desktop users will probably have more cores than parititons; sorting out a search from couple/few partitions sounds like a software limitation?
TomWomack - Friday, November 13, 2020 - link
That is a $7000 FPGA which is 3/4 filled with the logic to drive the PCIe and SSD interfaces, which makes the discussion of 'servers with dozens of them in' quite alarmingly expensive; also leads me to wonder why they've integrated it with $800 worth of flash. I am also quite surprised not to see a QSFP+ port on the SSD. I'm sure there is some really exciting use case that they will write a white paper about, but this is really very esoteric for the front page of anandtech.AlexDaum - Friday, November 13, 2020 - link
I'm sure Samsung can get that FPGA cheaper when they buy large quantities, but it's still going to be expensive.MrSpadge - Friday, November 13, 2020 - link
Well, that 7000$ is surely not the fabrication cost, just the end user price, right?SaberKOG91 - Friday, November 13, 2020 - link
BittWare make an accelerator called the 250-U2 which seems to retail for ~2800$ with 8GB of DDR4. They also have a PCI-E card version with 3.84TB of storage which goes for ~6600$. Samsung cuts out the overhead by selling their own Flash, so I wouldn't be surprised if MSRP was closer to 5-6k$.I'm not at all surprised there's no QSFP+. U.2 is a great form-factor for both compatibility with existing storage server designs and handling the thermals.
That said, people seem to be missing the point of these devices. If you operate on large datasets, being able to optimize access at the device level can significantly reduce the amount of computing power needed to do the same operations with a CPU in RAM, especially when you are using an FPGA to implement parts of the hardware. The cost of one of these easily outweighs the cost of having multiple NVMe drives to satisfy the same I/O requirements and then having to rely on sub-optimal CPU algorithms for processing that data. This will be a huge boon to database and search engine performance. You could also use them to implement rather large content-addressable memories for networking applications and the like.
Spunjji - Monday, November 16, 2020 - link
"That is a $7000 FPGA which is 3/4 filled with the logic to drive the PCIe and SSD interfaces"The SSD interface doesn't run on the FPGA - just the PCIe switch. There's no indication that it's "3/4" full, and in fact that seems like a bizarre assumption to make given the obvious downsides that you note.
Tomatotech - Friday, November 13, 2020 - link
That feeling when a hard drive has more computing power than every computer you own ...silverj42 - Friday, November 13, 2020 - link
Link to buy one, please?PeachNCream - Friday, November 13, 2020 - link
Thinking this would be a great place for bored systems administrators to hide low end gaming hardware on corporate servers. One more thing we have to figure out how to monitor and secure also so its certainly a risk-vs-reward critter.Spunjji - Monday, November 16, 2020 - link
It would be so, so much cheaper (and faster) just to throw a GPU in there and come up with an ad-hoc rationalisation for it.Rοb - Friday, November 13, 2020 - link
Here are the product links: https://samsungsemiconductor-us.com/smartssd/ --- https://www.xilinx.com/applications/data-center/co... --- https://www.mouser.com/ProductDetail/Xilinx/XCKU15...There's a "How to buy" link, but it presumes that you know "how to use" (you would want to have experience programming FPGAs). 4TB seems to be a rather tiny offering, but there are already Cloud Services that you can test out these drives.
There are a lot of use cases listed on the above links, but if you can't justify U$7K for a very smart 4TB of storage this wasn't made for you; try the Cloud Services already setup.
silverj42 - Friday, November 13, 2020 - link
So 6k euros for 1.1M LE + 4GB ram + 4TB storage. Quite expensive. The NiteFury I am using now was 300 euros for 200k LE + 512MB ram. Perhaps this is worth it, actually. Thank you for the mouser link!TomWomack - Saturday, November 14, 2020 - link
The Mouser link is for a single FPGA chip rather than for the whole SmartSSD device discussed here.silverj42 - Saturday, November 14, 2020 - link
Oh, thank you. Saved me that 6k :-)silverj42 - Friday, November 13, 2020 - link
> There are a lot of use cases listed on the above links, but if you can't justify U$7K for a very smart 4TB of storage this wasn't made for you; try the Cloud Services already setup.I read this as someone is already renting a cloud instance with one of these installed. As you seem to conjure up links I could not find, is there a link for this service as well? Does Hetzner have these or ?
Billy Tallis - Friday, November 13, 2020 - link
Nimbix has a page up about having SmartSSDs in their public cloud, but it doesn't appear they've added that option to their price calculator: https://www.nimbix.net/samsungsmartssdsilverj42 - Saturday, November 14, 2020 - link
Thank you - so not available, and the SSD logic takes 800k out of the 1.1M LEs, so this is basically not a real contender for any price above $1000 for me. Too bad, it seemed great at a first glance.Billy Tallis - Saturday, November 14, 2020 - link
LEs aren't the same as LUTs. The FPGA has 523k LUTs, of which about 300k LUTs are available for accelerator IP—roughly equivalent to a KU11P.erinadreno - Saturday, November 14, 2020 - link
Samsung should use ASIC PCIe switch rather than using LUTs on the FPGA. FPGA is for programming, the PCIe switch is a given. But I guess Xilinx just want user to design some DSP only algorithm, which is not totally unreasonable.Billy Tallis - Saturday, November 14, 2020 - link
There aren't a lot of options for small 12-lane PCIe switches. The PLX PEX8714 is 19x19mm and 2.7W, and ASM2812 is 21x21mm. I'm not sure how easy it would be to squeeze one of those into an already crowded U.2 drive.eastcoast_pete - Wednesday, November 18, 2020 - link
It's cool tech for sure, but I am not sure I get why this has to be implemented as an FPGA, and not exactly a small and cheap one, either? Could this be done using, let's say, in software with a decently fast CPU/NPU combo? Why does it have to be implemented as field-programmable hardware?