Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us $3,400 per month for bandwidth bills alone, and since we don't believe in shoving popup ads to our registered users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
spunkshui
Oct 5, 2011





Klyith posted:

I just cannot bring myself to care a single iota about GPUs being resold, scapled, price hiked, secretly ebayed by OEMs, or anything else. We're not talking lifesaving drugs here. They're unnecessary products for a nerd hobby. If nvidia couldn't produce enough of them to satisfy demand then the pricetag is naturally going to trend up to whatever the biggest richest idiots are willing to pay. The best solution is to not be a big idiot and just wait.


(I will also note that we're now 2/2 for hosed up hardware at launch from nvidia cards -- the 2000 space invaders corruption, and the 3000 bad capacitors. So lol at all the people who bought 3080s before even knowing what the potential competition was, with a decent segment of them not even caring what Big Navi might be because "AMD drivers bad".)

1. Gpus are used for more then games.
2. If MSI is funneling cards like this they are breaking Nvidia rules. The reason we dont have any cards for $3000 at bestbuy is nvidia control.
3. The caps are not bad.

Adbot
ADBOT LOVES YOU

priznat
Jul 7, 2009

Let's get drunk and kiss each other all night.

Big Navi/Big Vega

Some Goon
Jan 6, 2013

A golden helix streaked skyward from the Helvault. A thunderous explosion shattered the silver monolith and Avacyn emerged, free from her prison at last.




spunkshui posted:

1. Gpus are used for more then games.
2. If MSI is funneling cards like this they are breaking Nvidia rules. The reason we dont have any cards for $3000 at bestbuy is nvidia control.
3. The caps are not bad.

Best Buy could sell cards for $3000 all day if they wanted, MSRPs aren't enforceable.

spunkshui
Oct 5, 2011





Some Goon posted:

Best Buy could sell cards for $3000 all day if they wanted, MSRPs aren't enforceable.

They would then lose their ability to sell the product.

And if MSI keepes doing this poo poo you better believe that they would lose the ability to be a board partner.

The reason we have a good price at all is because of Nvidia doing this poo poo.

If Nvidia didn’t do it none of the cards would be under $1000 at any store whatsoever.

This is a good thing. It’s crazy to think that Nvidia shouldn’t do this or that it doesn’t need to be done.

spunkshui
Oct 5, 2011





Something tells me Best Buy is not eager to kill its entire computer department while it’s looking to be one of the few computer hardware big box retailers in America.

You are not going to see Best Buy charge $3000 for a gpu.

Nor are you going to see Best Buy open up an eBay account and start selling the cards for $3000 there.

MSI is probably trying to prove to Nvidia that’s not what they did right loving now.

Some Goon
Jan 6, 2013

A golden helix streaked skyward from the Helvault. A thunderous explosion shattered the silver monolith and Avacyn emerged, free from her prison at last.




No, they're not, but it's still a different relationship than between MSI and Nvidia.

Klyith
Aug 3, 2007

GBS Pledge Week


spunkshui posted:

1. Gpus are used for more then games.
2. If MSI is funneling cards like this they are breaking Nvidia rules. The reason we dont have any cards for $3000 at bestbuy is nvidia control.
3. The caps are not bad.

1. I'm fairly confident the availability of professional products for professional purposes has been good this whole time
2. Probably, that's between MSI and nvidia. But make no mistake, Nnidia does not care that GPUs are ending up on ebay in general. The restricted supply launch driving FOMO is an actual tactic in sales now.
3. the caps are underspecced / a bad design choice / a gently caress Up


spunkshui posted:

The reason we have a good price at all is because of Nvidia doing this poo poo.

The reason you have crazy out-of-whack demand is Nvidia taking enormous profit by ramping up prices for previous gen when they had weak competition, and previously soaking buttminers for everything they could get. They are lowering prices now because they see potential competition in play.

I don't think nvidia is bad for doing that, mind you. Capitalists gonna capitalize. But they bought ARM for 40 billion dollars, and they didn't get that money by giving away GPUs. They got $40 billion from our collective wallets, and the people with the least patience and cost/benefit neutrality contributed the most money.

spunkshui
Oct 5, 2011





Klyith posted:

3. the caps are underspecced / a bad design choice / a gently caress Up

Did I miss something in this?

The caps work fine. The TUF card had the same crash and it had all the "better caps."

https://twitter.com/HardwareUnboxed...4468298753?s=20

Subjunctive
Sep 12, 2006

careful now


Klyith posted:

But they bought ARM for 40 billion dollars, and they didn't get that money by giving away GPUs. They got $40 billion from our collective wallets

Really? I though the majority of that deal was stock, and that they “only” had 10-15B in cash.

Klyith
Aug 3, 2007

GBS Pledge Week


spunkshui posted:

Did I miss something in this?

The caps work fine. The TUF card had the same crash and it had all the "better caps."

https://twitter.com/HardwareUnboxed...4468298753?s=20

Beats me, I'm not constantly on hardware twitter. Are we at least sure that there's a major difference between the founder's cards and the partners? Because that would be even more Not Nice, probably it would indicate that the 3080's stock spec is a bit too ambitiously set and that only the founder's, which are reportedly cherry-picked by nvidia at the fab, are truly stable at the 'stock' speed.


Subjunctive posted:

Really? I though the majority of that deal was stock, and that they “only” had 10-15B in cash.

Their stock is pretty correlated to how many video cards they sell.

Plus the whole AI / ML thing. I suppose a computer program that plays go really well probably contributed about $10 billion out of that 40, which is pretty funny when you think about it.

Subjunctive
Sep 12, 2006

careful now


Klyith posted:

Their stock is pretty correlated to how many video cards they sell.

Give or take the irrationality of the stock market, yeah. I’d apparently misunderstood you saying that they paid for ARM out of GPU revenues, and $40B of accumulated nerd-tithe savings seemed like a lot more than I thought they had on hand.

spunkshui
Oct 5, 2011





Klyith posted:

Beats me, I'm not constantly on hardware twitter. Are we at least sure that there's a major difference between the founder's cards and the partners?

No.

This thing you keep saying is happening is not happening.

edit: unless i'm really really wrong, I would love for you to give info where you got this

Klyith
Aug 3, 2007

GBS Pledge Week


spunkshui posted:

No.

This thing you keep saying is happening is not happening.

edit: unless i'm really really wrong, I would love for you to give info where you got this

What part is not happening, the crash to desktop in games? It was definitively happening and multiple reports said that capacitor choices were at least part of the issue. If capacitors were only part of the issue then that's fine. But still it is a thing that happened. Games were crashing, the launch has had issues beyond supply.

The chips being binned? Only repeating what others have said, but it was something they absolutely did for the 2000 launch. "A" chips with better performance were sold for more and reserved for factory-OCed and the founders cards.

spunkshui
Oct 5, 2011





Klyith posted:

What part is not happening, the crash to desktop in games? It was definitively happening and multiple reports said that capacitor choices were at least part of the issue.

The crash also effects the TUF at stock speeds.

How is that a “capacitor choice issue.”

?

spunkshui fucked around with this message at 00:28 on Oct 8, 2020

Subjunctive
Sep 12, 2006

careful now


Klyith posted:

What part is not happening, the crash to desktop in games? It was definitively happening and multiple reportscases of unsubstantiated speculation said that capacitor choices were at least part of the issue.

Klyith
Aug 3, 2007

GBS Pledge Week



When EVGA went and gave statements saying that capacitors were a problem, it's much more than speculation by randos. Like, good that it turned out not to be the case, and I apologize for repeating it when I'd been out of the loop on the issue.

Anyways, launch 3080s have CTDs for mystery reasons, so I'll go back to the original point that getting FOMO about day 1 launch hardware is maybe a bad choice.

K8.0
Feb 26, 2004

Her Majesty's 56th Regiment of Foot

Every loving GPU crashes at launch jesus christ. I guarantee you whatever GPU AMD is launching is going to crash more than the 3080, and it will take far longer to fix.

CaptainPsyko
May 2, 2004

We're Gonna Run.

We're Gonna Crawl.

Kick Down Every Wall.


Klyith posted:

When EVGA went and gave statements saying that capacitors were a problem, it's much more than speculation by randos. Like, good that it turned out not to be the case, and I apologize for repeating it when I'd been out of the loop on the issue.

Anyways, launch 3080s have CTDs for mystery reasons, so I'll go back to the original point that getting FOMO about day 1 launch hardware is maybe a bad choice.

EVGA's statement was accurate for the specific design of their own cards, for performance reasons relative to the targets they set or whatever, but the CTD issue had literally zero to do with capacitor-gate at the end of the day, and was fixed by a driver update; it affected FE cards, ASUS TUF cards that had a full array of "the good caps", and everything in between. One issue wasn't correlated to the other in any way.

Subjunctive
Sep 12, 2006

careful now


Klyith posted:

When EVGA went and gave statements saying that capacitors were a problem, it's much more than speculation by randos.

Wasn't EVGA's statement about hitting a boost target and not about crashing?

e;fb

Subjunctive fucked around with this message at 01:13 on Oct 8, 2020

Wirth1000
May 12, 2010

the best driver
the best team
forza vettel


https://www.youtube.com/watch?v=V0JuW6x8lsw

Why do I feel someone already did this recently? I mean it's not exactly the most original idea to begin with.

Wirth1000
May 12, 2010

the best driver
the best team
forza vettel


Oh, yeah....

https://www.youtube.com/watch?v=eQ_8F4nzyiw

That's what I was thinking of.

Mu Zeta
Oct 17, 2002

Me crush ass to dust


Grimey Drawer

Hey the job is called "content creator" and not artist or journalist.

SCheeseman
Apr 23, 2003



The best bang for your buck PCs you can get from places like aliexpress are based on old Xeon stuff, using "new" motherboards that contain salvaged server parts. They aren't great and single thread performance kinda sucks, but there's usually enough cores to make up for it and a lot of modern games run fine on them. They're only really useful to people living in countries where getting new, value oriented PC parts is cost prohibitive, though.

Endymion FRS MK1
Oct 28, 2011



Wirth1000 posted:

Are they actually decent? All the photos and videos of them kinda made them seem pretty cheap. On par with your typical low cost house brand toolsets from big box hardware stores.

The printing of the labels on the drivers were kinda trash but I think they fixed it in newer runs? The tools themselves are solid and feel great

Mr.Radar
Nov 5, 2005

You guys aren't going to believe this, but that guy is our games teacher.


SCheeseman posted:

The best bang for your buck PCs you can get from places like aliexpress are based on old Xeon stuff, using "new" motherboards that contain salvaged server parts. They aren't great and single thread performance kinda sucks, but there's usually enough cores to make up for it and a lot of modern games run fine on them. They're only really useful to people living in countries where getting new, value oriented PC parts is cost prohibitive, though.

Yeah, PhilsComputerLab covers a lot of these. Here's a recent video he made on how viable the "X58" platform is these days:

https://www.youtube.com/watch?v=nO7aNUlTlaE

Theophany
Jul 22, 2014

"This is only the beginning of the reckoning. This is only the first sip, the first foretaste of a bitter cup which will be proffered to us year by year unless, by a supreme recovery of moral health and martial vigour, we arise again and take our stand for freedom as in the olden time."


Wirth1000 posted:

https://www.youtube.com/watch?v=V0JuW6x8lsw

Why do I feel someone already did this recently? I mean it's not exactly the most original idea to begin with.

I really want that case and always have since I was like 14.

Fame Douglas
Nov 20, 2013

RELY NOT ON MY HONOR!!! FOR WHEN I OFFER MY WORD OF BOND, I TAKE NOT THAT VOW TO HEART!! CASUALLY, I BRING SHAME TO MY HOUSEHOLD AND RUIN TO THOSE WHO RELY ON MY COMMITMENT, BY SHIRKING MY AVOWED DUTY

K8.0 posted:

Every loving GPU crashes at launch jesus christ. I guarantee you whatever GPU AMD is launching is going to crash more than the 3080, and it will take far longer to fix.

Funnily enough, my 2080 Super started crashing as well with the latest release driver (the new hotfix driver available via their support site fixes it).

VorpalFish
Mar 22, 2007
reasonably awesometm

All the capacitor stuff was just people repeating and amplifying speculation from igors lab or whatever.

As it turns out there's already a driver fix with no performance impact so it looks like the whole cap thing was much ado about nothing.

EngineerJoe
Aug 8, 2004
-=whore=-

VorpalFish posted:

All the capacitor stuff was just people repeating and amplifying speculation from igors lab or whatever.

As it turns out there's already a driver fix with no performance impact so it looks like the whole cap thing was much ado about nothing.

EVGA did corroborate the findings saying they discovered the all 'poscap' design didn't pass their testing.

Subjunctive
Sep 12, 2006

careful now


EngineerJoe posted:

EVGA did corroborate the findings saying they discovered the all 'poscap' design didn't pass their testing.

That was about it not reaching their target boost clocks, not about it crashing, I believe.

repiv
Aug 13, 2009



College Slice

Subjunctive posted:

That was about it not reaching their target boost clocks, not about it crashing, I believe.

The exact wording was that the all POSCAP layout "cannot pass the real world applications testing" which doesn't make it clear whether they meant crashing or something else

VorpalFish
Mar 22, 2007
reasonably awesometm

I mean you had mlc cards crashing before and it seems like all spcap designs are fixed with the new firmware so...

At most it seems like the mlc caps might get you like 10-40mhz more clockspeed.

K8.0
Feb 26, 2004

Her Majesty's 56th Regiment of Foot

There are boatloads of other MLCCs (including some that have spots but are not by spec filled on the reference design) so the concept of "All MLCC" vs "All POSCAP" was a complete misunderstanding / misstatement of the situation to begin with. It may well be true that for the particular setup EVGA chose, an "All POSCAP" design did not work. But it's quite clear that for some other implementations where other parts of power delivery and filtering were done differently, it works just fine. Buildzoid (Actually Hardcore Overclocking) did a pretty good video about it, as did some other people.

movax
Aug 30, 2008



K8.0 posted:

There are boatloads of other MLCCs (including some that have spots but are not by spec filled on the reference design) so the concept of "All MLCC" vs "All POSCAP" was a complete misunderstanding / misstatement of the situation to begin with. It may well be true that for the particular setup EVGA chose, an "All POSCAP" design did not work. But it's quite clear that for some other implementations where other parts of power delivery and filtering were done differently, it works just fine. Buildzoid (Actually Hardcore Overclocking) did a pretty good video about it, as did some other people.

I don't know if it is as much of a thing in a lot of the more modern multi-phase regulator designs, but engineers can sometime get bitten with regulator instability with too low ESR (equivalent series resistance) at the output of a DC/DC, as that term ends up creating a zero for the control loop (mathematical zero). If it's too low (more of a problem on older ones), it can actually cause a problem, so you need to add a tantalum or similar with higher ESR to get that zero back.

The basics of PDN design I guess now are a YouTuber thing, but... different tools do different things. MLCCs even are split into Class I / Class II dielectrics that trade stability for volumetric efficiency (i.e. 18 pF C0G/NP0 caps you'd use for an oscillator load are Class I, 10 uF X7Rs are Class II). All about what you're trying to do with the PDN which is generally keep the impedance across frequencies below a target value. Little bit of knowledge is a dangerous thing — I wouldn't be surprised if the PDN designs are designed around being tolerant to some caps not actually being populated correctly (production compensation) and questionable quality capacitors that have far worse DC Bias than they actually say.

FuturePastNow
May 19, 2014




lol some nerds are salty at the Retro Man Cave guy rebranding his channel to take Man out of the name to be nice and inclusive

Adbot
ADBOT LOVES YOU

Kingnothing
Aug 2, 2007




FuturePastNow posted:

lol some nerds are salty at the Retro Man Cave guy rebranding his channel to take Man out of the name to be nice and inclusive

https://youtu.be/WLMgbV3uaz8

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply