Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

RaptorCS's redemption: the POWER9 machine works

drewdevault.com

Post reply on HN
Print view View original

41–50 of 60 posts

Re: RaptorCS's redemption: the POWER9 machine works

#41

Post by classichasclass »

rjsw wrote: post #39
classichasclass wrote: post #35

I hope this means some future work on Wayland performance on ppc64le or at least on 2D only is coming. My GPU-less 4-core Blackbird was unusable in Wayland with the basic BMC graphics (whereas it is quite sprightly in X.org on the same version of Fedora and with the exact same hardware loadout). I'll be honest and say I'm a Wayland skeptic generally for reasons I won't derail this post with, but I'm resigned to it pr…

What is stopping people from adding a GPU to this class of system ?

From a hardware perspective, nothing. I have one in my Talos II and it runs rather better in Wayland than my Blackbird, though I'm usually in X.org. My only objection here is why performance would be so terrible without one. It was also part of an experiment to see how cheap I could make the loadout and still have an adequate system.

However, from a libre computing perspective, the WX7100 in my Talos II requires firmware which is non-open, even though it is included in Linux. I imagine that this market would include people interested in reducing the amount of firmware they don't control as much as possible, so for this niche, that is relevant.

Re: RaptorCS's redemption: the POWER9 machine works

#42

Post by emersion »

classichasclass wrote: post #35

I hope this means some future work on Wayland performance on ppc64le or at least on 2D only is coming. My GPU-less 4-core Blackbird was unusable in Wayland with the basic BMC graphics (whereas it is quite sprightly in X.org on the same version of Fedora and with the exact same hardware loadout). I'll be honest and say I'm a Wayland skeptic generally for reasons I won't derail this post with, but I'm resigned to it pr…

Which compositor are you using? If you don't have a GPU you'll want one with software rendering support. Using a software OpenGL implementation like llvmpipe will work but will be very slow.

We don't support this yet in wlroots, but it's definitely something we want to add, and not a limitation of the Wayland protocol at all.

Re: RaptorCS's redemption: the POWER9 machine works

#43

Post by tpearson-raptor »

qubex wrote: post #37
tpearson-raptor wrote: post #33

Earlier quoted context omitted.

For what it's worth, we do have a focus on making sure the hardware works properly etc.; I've always been pushing for that and haven't had any real problems there. This particular board was a test escape. A very unique, thankfully singular (so far as I know) test escape, but there's always a chance for that -- the larger failure in my mind was the breakdown in the support process. With our new controls in place (incl…

That particular horror story led me to not equip our development team with a dozen POWER9 machines, which is a pity, because the combination of almost-total trustworthiness and enormous firepower was extremely attractive for our particular niche (plus, I used to hand-code PPC64 AltiVec “velocity engine” code for HFT funds back when Apple sold G5 Xserves). I’m glad to hear that it was a one-off and that you have revis…

It also won't happen at this point, ever. Is there anything we can do to help you revisit this decision?

Re: RaptorCS's redemption: the POWER9 machine works

#44

Post by wahern »

zwegner wrote: post #36
tpearson-raptor wrote: post #28

Earlier quoted context omitted.

I might take that bet core for core, matched node size (i.e. 14nm), run out of tmpfs or NVMe storage, and with both systems protected against Spectre, Meltdown etc. (i.e. no cheating by ignoring the ISA specifications)... Slightly OT, but if you limit the playing field to open ISA / fully owner controlled systems, POWER9 is so far out front in terms of performance that it looks like an outlier. ;)

> both systems protected against Spectre, Meltdown etc. (i.e. no cheating by ignoring the ISA specifications) Does the x86 specification say "speculation will have no observable side effects on the memory subsytem"? I wasn't aware of that. In other words, Spectre/Meltdown are bad security flaws, but I don't think they're the result of cheating.

Meltdown is absolutely cheating as it speculates through permission boundaries. I don't see how anyone could argue otherwise in good faith, especially considering that it was only Intel and IBM effected. AMD and SPARC were immune, and the extent of ARM's vulnerability was a single register, which seems like a bug, not reflective of an architectural design decision.

You can't design your software to protect against Meltdown, which makes the ISA guarantees useless. Whereas Spectre-type side channels can be mitigated through software, which is what cryptographic algorithm implementations have been doing for years.

There's far more room for debate regarding culpability for Spectre class attacks, though it's pretty clear that Intel deliberately pushed the envelope in ways that AMD and ARM weren't prepared to do.

Re: RaptorCS's redemption: the POWER9 machine works

#45

Post by bjoli »

kop316 wrote: post #11
Sir_Cmpwn wrote: post #9

Earlier quoted context omitted.

Not a dumb question. https://sourcehut.org should tell you what you need to know. If you have any specific questions, I can answer them.

Hello! I went through that website before posting. That's why I thought it was a Gitlab-like service. I like that it works without JavaScript, and it appears to have a wiki-like integration too. It looks like a neat service.

I migrated after bitbucket decided to kill hg support. I could probably get by using git, but I like that I don't have to :)

All in all, I have been very happy.

Re: RaptorCS's redemption: the POWER9 machine works

#46

Post by qubex »

tpearson-raptor wrote: post #43
qubex wrote: post #37

Earlier quoted context omitted.

That particular horror story led me to not equip our development team with a dozen POWER9 machines, which is a pity, because the combination of almost-total trustworthiness and enormous firepower was extremely attractive for our particular niche (plus, I used to hand-code PPC64 AltiVec “velocity engine” code for HFT funds back when Apple sold G5 Xserves). I’m glad to hear that it was a one-off and that you have revis…

It also won't happen at this point, ever. Is there anything we can do to help you revisit this decision?

Nope, too late. The decision was made and orders have been placed. You don’t get do-overs, sorry.

I might still get a machine for my own amusement, but... I don’t know, reading of some poor guy left dangling for best part of a year just doesn’t build confidence and leaves me with a “there but for the Grace of God go I” feeling, you understand?

Re: RaptorCS's redemption: the POWER9 machine works

#47

Post by classichasclass »

emersion wrote: post #42
classichasclass wrote: post #35

I hope this means some future work on Wayland performance on ppc64le or at least on 2D only is coming. My GPU-less 4-core Blackbird was unusable in Wayland with the basic BMC graphics (whereas it is quite sprightly in X.org on the same version of Fedora and with the exact same hardware loadout). I'll be honest and say I'm a Wayland skeptic generally for reasons I won't derail this post with, but I'm resigned to it pr…

Which compositor are you using? If you don't have a GPU you'll want one with software rendering support. Using a software OpenGL implementation like llvmpipe will work but will be very slow. We don't support this yet in wlroots, but it's definitely something we want to add, and not a limitation of the Wayland protocol at all.

It was whatever was out of the box with Fedora 30 and GNOME (I guess gnome-shell itself?), and indeed was very slow with llvmpipe. I look forward to the future work on this.

Re: RaptorCS's redemption: the POWER9 machine works

#48

Post by qubex »

tpmx wrote: post #40
qubex wrote: post #37

Earlier quoted context omitted.

That particular horror story led me to not equip our development team with a dozen POWER9 machines, which is a pity, because the combination of almost-total trustworthiness and enormous firepower was extremely attractive for our particular niche (plus, I used to hand-code PPC64 AltiVec “velocity engine” code for HFT funds back when Apple sold G5 Xserves). I’m glad to hear that it was a one-off and that you have revis…

If I were you I'd be a lot more comfortable pulling the trigger after reading all of this. Being worried over what might have been spoilt milk if you had gotten a broken board in the past seems counterproductive.

The new machines I eventually decided to go with were ordered more than a week ago.

Re: RaptorCS's redemption: the POWER9 machine works

#49

Post by tpearson-raptor »

qubex wrote: post #48
tpmx wrote: post #40

Earlier quoted context omitted.

If I were you I'd be a lot more comfortable pulling the trigger after reading all of this. Being worried over what might have been spoilt milk if you had gotten a broken board in the past seems counterproductive.

The new machines I eventually decided to go with were ordered more than a week ago.

Sorry to hear that. If you ever get tired of worrying about untrustworthy firmware phoning home, DRM, etc., please consider us for the replacement purchase? ;)

Re: RaptorCS's redemption: the POWER9 machine works

#50

Post by tpearson-raptor »

qubex wrote: post #46
tpearson-raptor wrote: post #43

Earlier quoted context omitted.

It also won't happen at this point, ever. Is there anything we can do to help you revisit this decision?

Nope, too late. The decision was made and orders have been placed. You don’t get do-overs, sorry. I might still get a machine for my own amusement, but... I don’t know, reading of some poor guy left dangling for best part of a year just doesn’t build confidence and leaves me with a “ there but for the Grace of God go I ” feeling, you understand?

I understand. I also know this happened to no one else, and it was a pathological case in a support system that has since been completely overhauled. This is one reason why we started the self-serve RMA process -- to show that yes, we are committed to replacing defective hardware in a timely fashion. If the machine is defective, and we can't help via our online troubleshooting guide and Email / phone support, send it right back for replacement.

That being said, the assembled systems products are checked with a significant burn-in test etc. before they ship. We have never had an assembled system arrive defective anywhere, except if it was damaged in shipping (i.e. crushed box etc.).

Post reply on HN