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

We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

erikmcclure.com

Post reply on HN
Print view View original

231–240 of 362 posts

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#231

Post by Taniwha »

smilespray wrote: post #219
Taniwha wrote: post #211

Earlier quoted context omitted.

SuperMac - the first one was the "SuperMac Spectrum/24 series III", I can't remember the name of the big FPGA thing I built the accelerator on the last project (that 200 pin chip - look for a chip SQD-01 aka 'SQuiD') that was the basis of a whole family of subsequent cards (starting with the Thunder cards)

I loved those cards. My daily driver was a fully stacked IIfx with a SuperMac card. I remember telling my friends my display was more than a thousand pixels (1152 or something) wide.

yup 1152x870 - essentially the largest number that is roughly a 4:3 aspect ratio and less than 1 megapixel (2^20 VRAM was so very expensive) (and meets a bunch of other limitations like video dot clocks, VRAM clock rates, must be a multiple of the width of the VRAM banks etc)

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#232

Post by tcmb »

Hamuko wrote: post #198
ascar wrote: post #186

I'm surprised this isn't mentioned here more clearly: Some high end cases like the be quiet silent base I'm using have the option to mount the graphics card vertically, basically parallel to the mainboard, in a separate anchor slot. It needs an additional connector cable (~$20), but other than that is easy to setup, looks better with a look-in case (the illuminated fans face the glass side) and the weight pulls on a…

I like my mini-ITX case (Phanteks Shift Air 2) because it hangs the GPU from the top and it's sitting behind a mesh so it still has ample access to fresh air. https://www.kitguru.net/wp-content/uploads/2020/12/2-Still-8...

I'm getting a Cloudflare error page on that link: 'The owner of this website does not allow hotlinking to that resource'

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#233

Post by numpad0 »

One of the issues with that second bracket is there are not many cases that supports it. Mac Pro, HP workstations and desktop servers, and old, cheap looking, bent metal “DOS/V” cases has it, but modern “EATX” cases often don’t. It also cannot support too much weight.

Indeed, PCI standards were for adding interfaces to personal desktop computers after all. It does seem ill suited to host 450W cooperative sub-computers.

A more common approach to heavy expansion card is VME style chassis design. Off top of my head, NeXTcube, NEC C Bus, Eurocard uses this arrangement in consumer space, and many blade servers enclosures, carrier grade routers, and military digital equipments employ similar designs as well.

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#235

Post by Lramseyer »

pvarangot wrote: post #160
Lramseyer wrote: post #89

> Should we have GPU VRAM slots alongside CPU RAM slots? Is that even possible? I chuckled a little at this because I used to wonder the same thing until I had to actually bring up a GDDR6 interface. Basically the reason GDDR6 is able to run so much faster is because we assume that everything is soldered down, and not socketed/slotted. Back when I worked for a GPU company, I occasionally had conversations with co-wor…

I had the misfortune of bringing up firmware for an overengineered piece of crap with DDR5 and even though we had the proper measuring equipment instrumenting properly is barely possible. It's designed with reflections in mind and there's a loopback channel at a slower word rate that I think the controller uses to characterize the analog signal because it's like, barely digital.

Did Synopsys design the DDR5 controller by any chance?

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#236

Post by Godel_unicode »

hinkley wrote: post #153
Lramseyer wrote: post #89

> Should we have GPU VRAM slots alongside CPU RAM slots? Is that even possible? I chuckled a little at this because I used to wonder the same thing until I had to actually bring up a GDDR6 interface. Basically the reason GDDR6 is able to run so much faster is because we assume that everything is soldered down, and not socketed/slotted. Back when I worked for a GPU company, I occasionally had conversations with co-wor…

What if GPU makers made motherboards?

All the big ones do already.

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#237

Post by Godel_unicode »

zapdrive wrote: post #122
MBCook wrote: post #70

Earlier quoted context omitted.

My proposal isn’t too different. Move the one ultra-fast PCIe slot to the top of the motherboard. It would be mounted so the GPU plugs in parallel to the motherboard, on the same plane, above the motherboard on a tower case. The few other PCIe slots that exist can stay at the bottom. Only downside is the case has to be taller. Not sure if that would be considered a problem or not. This doesn’t really help dual GPU se…

On a further note, why does it even have to be inside the case? Make a slit in the case on the top, so that the PCIe slot is sticking out. Stick a GPU in that slot, supported by the case. The GPUs these days look much cooler anyways.

Briefly, EMI.

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#238

Post by TT-392 »

I think a start would be providing an extra screw hole in a standardized place, so that case manifacturers can design for this, instead of putting the card between 2 plastic clamps. A problem here would be length differences in cards though. But I think that it being in the middle of most long cards, and at the end of the smallest card isn't the craziest idea.

Re: We Need New Motherboards Before GPUs Collapse Under Their Own Gravity

#240

Post by Godel_unicode »

terom wrote: post #208
luminiferous wrote: post #72

Earlier quoted context omitted.

LTT has a video where they tried to see how many PCIe riser cables they could string together before it stopped working.[1] They got to several meters. Maybe you could argue that it's worse inside a PC case since there's more EMI, but it seems like your PCIe riser cable would have to be very out of spec before you'd notice anything. [1] https://www.youtube.com/watch?v=q5xvwPa3r7M

I wonder if that benchmark actually loaded the PCIe bus to any significant degree after the initial benchmark startup, or just updated a couple small parameters on a single scene and thus mainly just tested the local computation on the GPU? You'd want to somehow monitor the PCIe bus error rate - with a marginal signal and lots of errors -> retries, something that loads the bus harder (loading new textures etc) could…

You need Steve for that kind of testing, LTT would be busy putting rgb on it and then (badly) water cooling it so they could sell you a backpack with no warranty.
Post reply on HN