The next Raptor OpenPOWER systems are coming, but they won't be Power10
Raptor yesterday officially announced that we're not getting Power10 systems. The idea is we're going to be getting something better: the Solid Silicon S1. It's Power ISA 3.1 and fully compatible, but it's also a fully blob-free OpenPOWER successor to the POWER9, avoiding Power10's notorious binary firmware requirement for OMI RAM and I/O.
I asked Timothy Pearson at Raptor about the S1's specs, and he said it's a PCIe 5.0 DDR5 part running from the high 3GHz to low 4GHz clock range, with the exact frequency range to be determined. (OMI-based RAM not required!) The S1 is bi-endian, SMT-4 and will support at least two sockets with an 18-core option confirmed for certain and others to be evaluated. This compares very well with the Power10, which is also PCIe 5.0, also available as SMT-4 (though it has an SMT-8 option), and also clocks somewhere between 3.5GHz and 4GHz.
S1 embeds its own BMC, the X1 (or variant), which is (like Arctic Tern) a Microwatt-based ISA 3.1 core in Lattice ECP5 and iCE40 FPGAs with 512MB of DDR3 RAM, similar to the existing ASpeed BMC on current systems. X1 will in turn replace the existing Lattice-based FPGA in Arctic Tern as "Antarctic Tern," being a functional descendant of the same hardware, and should fill the same roles as a BMC upgrade for existing Raptor systems as well as the future BMC for the next generation systems and a platform in its own right. The X1 has "integrated 100% open root of trust" as you would expect for such a system-critical part.
Raptor's newest systems are planned for late 2024. There will be tiering, so most likely (though not confirmed) Blackbird, T2 and T2 server classes of systems will be available under new names. Price? Well, you'll just have to wait and see.
Solid Silicon is definitely a new name in the Power ecosystem and we don't know a lot about them. There's a web page, but the TwXitter and LinkedIn links are unpopulated as of this writing, and it's maddeningly minimal on actual content. Tim confirmed they are a new licensee and have been working on the design for at least a couple years. The press release gives a 737 area code, which is Austin, Texas, and the only Solid Silicon business entry I could find for Texas is this one for Solid Silicon Technology LLC in Plano. I'm told this isn't them, so if anyone from Solid Silicon would like to lift the corporate veil a little, drop me a line at ckaiser at floodgap dawt com. [UPDATE: The LinkedIn was updated after this posted, listing Todd Rooke as CEO. Rooke's listing indicates past experience with FPGAs, as well as his time at HPE and Microsoft. His location is given as Colorado Springs but Colorado lists no company by that name. Hopefully more to come.]
But besides new systems in the offing, it's also good news that we're getting — we hope — performant OpenPOWER chips that aren't from IBM. I don't have anything against IBM; I've worked with IBM hardware for literally decades, and my home server is a classic POWER6 that just keeps on truckin'. But IBM designs chips to benefit IBM's world, which is server rooms (ask anyone who's got one what it's like to share an office with a POWER8), and IBM doesn't do end-user sales. If Raptor has a good partner here who can design solid OpenPOWER chips for workstations and small servers, not traditionally IBM's present domain but one important for them to maintain if they want OpenPOWER to stay relevant, then in around a year we should be in for a treat — and a very rosy near future.
If you have a line to Raptor, can you find out if the new S1 is going to be bi-endian? Considering the official announcement said "direct upgrade", I'm hoping that means that OSes like Adélie Linux and OpenBSD will continue to function on the new chip.
ReplyDeleteWe still have no plans at Adélie to support little-endian Power in the near to mid-term. There's no reason to dedicate our limited resources to a new platform when big-endian runs old and new systems, and we've been able to bring up desktop support even on the modern systems. But with the latest rumblings, I have personally grown very concerned.
Bi-endian. I updated the article.
DeleteHello i really wanted power 10 cpu for new raptor computer. Bad news. : (
ReplyDeleteSorry for the novice question, but what's the effective differences between a new PowerISA 3.1 CPU design, and Power10? Does that just mean it's not an IBM chip? Can is simply not be called that?
ReplyDeleteRight, it just means it's not an IBM chip. The Power10 implements ISA 3.1, so the S1 would be compatible with it, but I imagine the new name is to avoid confusion and make it clear it's not saddled with the original's baggage.
DeleteThe most pressing question is, who is going to fab these chips, and at what node? This will have massive impact on both performance and power consumption as well as cost: And nodes that are competitive with what other architectures have on offer are expensive.
ReplyDeleteThe S1 cannot be anything other than a Power10. A no-name company is not capable of developing a CPU in such a short time frame.
ReplyDeleteRaptor has long assumed that they can use the Power10 next. Then Raptor learned that the Power10 will have blobs.
Then they asked Synopsis if there was a way to avoid the blobs, which Synopsis first confirmed and then later said, no we won't make it possible.
This means that Raptor was not aware until 2021 that they will not be able to use the Power10 (in this form).
And that a company like Solid Silicon was thinking in 2018, let's make a PowerPC CPU because there will be a market for it. No way!
I don't think anyone implied it was a from-scratch design. Indeed it cannot be if it's intended to be a competitive performer. Just what exactly it is, where it comes from, and how they're planning to get it made, is still not entirely clear from the announcements.
Deleteto me, the best case could be: if they licensed chunks of the power10 IP from IBM, and stripped out the bad bits, it's feasible that they could have licensed the memory controller or done it themselves, and then get the result fabbed at the same places power10 is done at, we would basically end up with what we wanted from the start
Deletecompletely idle speculation, anyway
You are right. We know next to nothing. Not even the manufacturing process. Only that there will be an announcement in the second half of 2024. That means if it was a Power10 design, there would be no reason to wait another 9 months.
DeleteSo there must be a reason why Raptor / Solid Silicon is still waiting so long. It could be that some ominous big company wants nothing to be announced today. Maybe this ominous big company will make an announcement themselves. Maybe something between February 18-22, 2024 ;-)
Anything else doesn't really make sense and Raptor is even right when they say it won't be a Power10. But a Power11 8-)
But it is stated to be Isa Version 3.1, which would be Power10 levels of features. Wouldn't Power11 have a newer version of the Isa ?
DeleteThat is correct. But I wouldn't be surprised if Solid Silicon corrects itself and says it's not PCIe 5.0, but 6.0 and not PowerISA 3.1, but 3.2. But as already mentioned, pure speculation.
DeleteIn the end, it will be an IBM Power design and nothing else.
Correct me if I’m wrong, but I think Power ISA 3.1 has been designed in an extensible way, which actually makes POWER9 also compliant with specification version 3.1, just without the new optional extensions that Power10 supports. So, assuming my understanding is correct, just the fact the ISA version is 3.1 does not even necessarily tell us what features are supported. I doubt the ambiguity was intended in the announcement, but if we want to make things even more uncertain, we totally can…
DeleteThis is actually what I was hoping for. IBM really hasn't been interested in workspace/small servers since before they sold all their x86 stuff off. While we have some info here that the BMC is based on the work done off of the MicroWatt and Raptor CS BMC, I really hope a bunch of the other work is based-on, and potentially gets upstreamed into MicroWatt and other repos. It'd be quite nice to have POWER that implements an ISA, a spec, and can be looked at to see the peculiarities of it. It'd also be nice to have a QEMU platform for it so we can start cobbling together software for it before it gets released.
ReplyDeleteGreat News. While the Lack of Information about Solid Sillicon is not Confidence inspiring, it could also be a stealth Startup. In any Case i am still hopefull about this.
ReplyDeleteOn Oct 4 i bought a blackbird from Raptor and it is still stuck on payment received. My support tickets and phone calls have been ignored, i also tried contacting Raptor engineering. There is a story of someone also ghosted by Raptor on the wiki who waited over a month for a update on there order, someone had to inform Raptor on Twitter about it. with little options left i am asking if someone with a line to raptor can help me(please). I want to like Raptor but this experience makes it vary hard, it feels like i have been robbed.
ReplyDeleteOnly three weeks? That is not yet time to panic ;-) Your ticket number?
DeleteEvery time I purchase something from raptor it takes months to hear back. I feel the same way and I don't know why I keep trying.
DeleteI was told the delay was from COVID burning through the area, so understaffed.
DeleteOct 4 guy here, my order has arrived. I believe that this new S1 design is keeping Raptor vary busy and there is no time for customer service. I tried to reply to people earlier but kept getting told to try again later.
DeleteWell, this is pretty good news! I figured this was based on libre-SoC's progress with Microwatt, if somewhere in the middle of its development, but then I saw that it's ISA 3.1... I mean, not that it's impossible to just add on the extra ISA stuff on top of 3.0b. I already have a P9 to stick in a Blackbird so I'm going to stick with it for a while, it's plenty good for what I need it for, but now I'm not sure whether to wait for release for people to offload theirs for cheap or buy it from Raptor to help fund the new systems. I'm leaning toward used just to reduce e-waste.
ReplyDeleteI'm happy to see it's coming either way, and that Raptor is doing better than I thought they were. It would also be neat if the new CPU also used the LGA 2601 socket and the chipset-equivalent could handle either CPU so that people can upgrade to the new chip on their own time, but I don't imagine it would.
If the schematics are correct, then the socket looks like one from AMD. Which would not be a new idea: https://www.theregister.com/2006/09/22/ibm_power7_opteron/
ReplyDeleteI would suggest looking at Delaware for companies by the name of Solid Silicon. I couldn't find anything, however.
ReplyDeleteWhat is interesting to note is that Raptor Eng. owns the Solid Silicon domain.
Hopefully Raptor takes over the “PowerPC Notebook” project as well and offers Si chip designs for that! https://www.powerpc-notebook.org/en/
ReplyDeleteThere is now an entry on Open Coporates. Based on the Ceo it must be them https://opencorporates.com/companies/us_co/20241491851 . According to this they are a branch of a company in Wyoming.
ReplyDelete