• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

AMD is a Software Company

erek

[H]F Junkie
2FA
Joined
Dec 19, 2005
Messages
12,839
De-prioritizing Hardware Initiatives

“Prominent AMD leadership, including Phil Guido, Executive Vice President & Chief Commercial Officer and Jack Huynh, Senior VP & GM, Computing and Graphics Business Group. AMD is making changes in a big way to how they are approaching technology, shifting their focus from hardware development to emphasizing software, APIs, and AI [virtualized] experiences.”

1720451567425.png


Source: https://www.techpowerup.com/324171/amd-is-becoming-a-software-company-heres-the-plan
 
Nvidia is where it is because of CUDA and the software platform they have created around it.
Nothing Nvidia is currently doing would be possible without it.
AMD and Intel both need to step up their software game to compete.
You can have the greatest hardware in the world and its useless unless there is software that can push it.
 
Software company with 4% and 6% operating margin the last 2 quarter in GAAP accounting.... under 50% in gross margin... that would be quite terrible number for a software company (gross tend to be 75%+, even a way more than software and a lot of hardware now company like Microsoft is at 70%).

I think the title should include the becoming
 
Nvidia is where it is because of CUDA and the software platform they have created around it.
Nothing Nvidia is currently doing would be possible without it.
AMD and Intel both need to step up their software game to compete.
You can have the greatest hardware in the world and its useless unless there is software that can push it.
Totally. It is about providing a meal, not ingredients.
 
Even simple hardware components often need drivers, which is software. My Corsair AIO cooler has firmware.

Cameras these days are just loaded with software, which they call firmware. And how about all those lines of code in the average new car. My dishwaher has firmware, but not upgradeable.
 
Even simple hardware components often need drivers, which is software. My Corsair AIO cooler has firmware.

Cameras these days are just loaded with software, which they call firmware. And how about all those lines of code in the average new car. My dishwaher has firmware, but not upgradeable.

Yeah, it has gotten out of hand.

1720467632186.png


(Though who the fuck grills on Thanksgiving?? Grilling is a summer only activity)

For me it has gotten to the point where I read the detailed specs of absolutely everything I buy these days. If there is any mention of "smart" or "wifi" or anything like that, I just move on. I won't buy it under any circumstance. (You know, actual computer/network/phone devices aside - of course)

I was in Home Depot the other day and in the lighting aisle they actually have bathroom fan/light combos with addressable RGB that accept voice commands.

Why the fuck would I ever want to talk to my bathroom fan?
 
Yeah, it has gotten out of hand.

View attachment 664196

(Though who the fuck grills on Thanksgiving?? Grilling is a summer only activity)

For me it has gotten to the point where I read the detailed specs of absolutely everything I buy these days. If there is any mention of "smart" or "wifi" or anything like that, I just move on. I won't buy it under any circumstance. (You know, actual computer/network/phone devices aside - of course)

I was in Home Depot the other day and in the lighting aisle they actually have bathroom fan/light combos with addressable RGB that accept voice commands.

Why the fuck would I ever want to talk to my bathroom fan?
So that rather than stumble around in the dark for the light and/or fan switch when you are drunk/hungover and badly needing to use the toilet, you can slur out for the lights and fan to come on rather than vainly seeking out the toilet, quit, and instead end up using the sink... poorly: leaving a foul smelling mess till morning.
 
Yeah, it has gotten out of hand.

View attachment 664196

(Though who the fuck grills on Thanksgiving?? Grilling is a summer only activity)

For me it has gotten to the point where I read the detailed specs of absolutely everything I buy these days. If there is any mention of "smart" or "wifi" or anything like that, I just move on. I won't buy it under any circumstance. (You know, actual computer/network/phone devices aside - of course)

I was in Home Depot the other day and in the lighting aisle they actually have bathroom fan/light combos with addressable RGB that accept voice commands.

Why the fuck would I ever want to talk to my bathroom fan?
The only "smart" devices I buy can operate normally without Internet.

For example I have smart light switches I automate, but they work perfectly fine as a normal light switch even without Internet. And they don't connect straight to the Internet, they connect to a hub I control locally.

My washer and dryer can give me phone notifications when they're done, but they still run perfectly fine without Internet. If you live somewhere like California it's only a matter of time before they mandate they're connected and control when you're allowed to use them.
 
The only "smart" devices I buy can operate normally without Internet.

For example I have smart light switches I automate, but they work perfectly fine as a normal light switch even without Internet. And they don't connect straight to the Internet, they connect to a hub I control locally.

Yeah, I used to feel that way, but honestly, I don't even trust them when they are disconnected these days, and I don't want to incentivize the market continuing to move in this direction, so no company will make any money off a "smart" or "wifi" enabled device from me.

Then there is the case of Unifi Video. They used to support a local server managing the video cameras, so I bought into the ecosystem. Then right as they announced the major hack of their servers a few years back, they discontinued the local Unifi video server and forced everyone onto the cloud.

I was like "fuck no", and disconnected everything I owned of theirs, and eventually migrated away from other products of theirs I used as well (like Unifi Wireless Access Points) because I simply don't trust them anymore.

Short of my PC and my phone, I want every last device I buy or use to be "dumb". I'm not going to reward companies with my money for this shitty market trend.

Yeah, I know. I'm a drop in the bucket compared to overall market trends, but I can't be the only one.
 
Totally. It is about providing a meal, not ingredients.
The ingredients matter to a meal. You would care if you found out that McDonalds used Pink Slime and the meat in their burgers are less meaty. AMD really needs to focus on their ingredients. Since AMD's problem is the ingredients, they need to get started on that. Especially when they depend on people like GeoHotz in helping them on GPU compute. I think that AMD lost their innovation when they fired Raja Koduri and have been playing catch up to Nvidia since. Even if AMD were to work heavily on software, they'd still need to innovate.
 
The ingredients matter to a meal. You would care if you found out that McDonalds used Pink Slime and the meat in their burgers are less meaty. AMD really needs to focus on their ingredients. Since AMD's problem is the ingredients, they need to get started on that. Especially when they depend on people like GeoHotz in helping them on GPU compute. I think that AMD lost their innovation when they fired Raja Koduri and have been playing catch up to Nvidia since. Even if AMD were to work heavily on software, they'd still need to innovate.
It is all about value to the customer. We're using abstract terms for parts, but a customer is trying to solve a problem. They don't care two squats where the problem is solved.
 
The ingredients matter to a meal. You would care if you found out that McDonalds used Pink Slime and the meat in their burgers are less meaty. AMD really needs to focus on their ingredients. Since AMD's problem is the ingredients, they need to get started on that. Especially when they depend on people like GeoHotz in helping them on GPU compute. I think that AMD lost their innovation when they fired Raja Koduri and have been playing catch up to Nvidia since. Even if AMD were to work heavily on software, they'd still need to innovate.
Perhaps but still, ROCm is supposedly AMD’s alternative to CUDA it launched in 2016, and 8 years in they are bragging about it working on a few select cards and being almost as good as CUDA for one or two small tasks that has a plethora of alternatives available. Meanwhile the current iteration of CUDA works on every Nvidia card going back to 2007.
It’s gonna run like ass depending on what your trying to do on those old cards but it will run, Ray Tracing included.

AMD needs to seriously work on their software and not be so reliant on “Open Source Contributor's”.
Yeah keep being open source but seriously put staff with an actual budget on the project not just a skeleton crew and a rough draft of a program and hope the community fills in the blanks for you.

Meanwhile Intel who’s only been in the graphics acceleration game a few short years is putting serious work into OneAPI which is steadily growing to be an actual CUDA alternative. Better yet it’s mostly platform agnostic, in that you can easily work with Intel or Nvidia hardware, and supposedly AMD. But sadly AMD is painful as ever because just getting ROCm to work on its own is an absolute nightmare let alone trying to get it to play with others.

Seriously I’m personally upset at AMD because I want a viable alternative to Nvidia in my servers I’ve got plenty of EPYCs and their great but their workstation and server GPUs are too rigid and the support scope is narrow to the point of being nonexistent.
I can spend a fortune on AMD hardware and make it work and then pray to the Omnissiah the scope doesn’t change during its operational lifecycle. Or I can buy Nvidia and throw it at whatever I need it to be and not only will it be supported (with the correct license $$$) but it will work damned well.

But I’m tired of signing those Nvidia invoices…
 
CUDA undeniably has been profitable for Nvidia and enabled them to be where they are today, but it has also - IMHO - been a huge problem for the market.

Proprietary standards and API's need to die. They are nothing but a way for a market leader to try to take over and completely dominate a market.

What Nvidia has done with CUDA since its inception should have been illegal.
 
Proprietary standards and API's need to die. They are nothing but a way for a market leader to try to take over and completely dominate a market.
But because of that possibility it create an incentive to put giant money and resource into them, you probably need to intervene and open them once it dominated for a while (a bit like patents have date on them).

Do the CUDA revolution happen if Nvidia cannot do it as a proprietary standards and are we better off ? how slower and different if they have to take a bunch of giant feedback and include people into it.

Balance should go toward openness for something like API, but maybe you do not want to completely destroy the reward for creating standard and api's...
 
From a business perspective this makes sense, I still think Nvidia has too much of a lead for them to surmount. But 5 years is a long time, AI could still be going strong or sales of AI chips could be coming down by that point who knows. If Nvidia is just kind of coasting off of blackwell sales and the next generation, it might give them a crack in the door to slip through and grab some market share. I still think they need to focus on hardware, particularly I think they should consider making ARM laptop chips if MS decides to continue to expand on ARM for windows. The M4 is miles ahead in terms of performance per watt, and it'd be great if there were an AMD solution to come close in that department as well.
 
The percentage of Nvidia employee that are multi-millionaire already (is it around 70% by now ?) could make it hard to stay hungry.

That said that talk from Huang about coming back to yearly release pattern with the current level of complexity and product whole stack, do not feel like coasting, it is aggressive, they are already talking about the Rubin platform that will double again most specs, but this time at a one year pace.
 
Dude, smoked turkey is amazing. Thinking about it, I grill/smoke year round but probably less in the Summer than the rest of the year. It's just too fucking hot.

Fair enough. I guess traditions differ based on region/climate.

To me Thanksgiving is a cold weather holiday. All about huddling indoors, and cooking robust cold weather dishes in the oven. I guess that might be different in places where late November is warmer.
 
But because of that possibility it create an incentive to put giant money and resource into them, you probably need to intervene and open them once it dominated for a while (a bit like patents have date on them).

Do the CUDA revolution happen if Nvidia cannot do it as a proprietary standards and are we better off ? how slower and different if they have to take a bunch of giant feedback and include people into it.

Balance should go toward openness for something like API, but maybe you do not want to completely destroy the reward for creating standard and api's...

Industries where working groups develop next gen API's work very well, and better stimulate competition in the marketplace.

And competition in the marketplace is after all what a capitalist market is all about. Maximizing competition in order to maximize benefit to the customer/society. That's why societies have competition law that outlaws many forms of monopoly and market manipulation.

Unfortunately these laws are rarely enforce harshly enough, which allows big organizations to consolidate, use their market leadership to achieve even more market dominance, and hurt customers, the market and society as a whole.

We - as a nation and a planet - need to get back to good old fashioned trust busting on a huge scale.
 
It is all about value to the customer. We're using abstract terms for parts, but a customer is trying to solve a problem. They don't care two squats where the problem is solved.
But when there's a problem we tend to point out the parts that aren't working. If I say my engine is ticking then you're not going to correct me and say the car is ticking. It's a terrible way to mask the real problem.
Perhaps but still, ROCm is supposedly AMD’s alternative to CUDA it launched in 2016, and 8 years in they are bragging about it working on a few select cards and being almost as good as CUDA for one or two small tasks that has a plethora of alternatives available. Meanwhile the current iteration of CUDA works on every Nvidia card going back to 2007.
It’s gonna run like ass depending on what your trying to do on those old cards but it will run, Ray Tracing included.
ROCm is a mess to the point where even I don't have it installed.
AMD needs to seriously work on their software and not be so reliant on “Open Source Contributor's”.
Yeah keep being open source but seriously put staff with an actual budget on the project not just a skeleton crew and a rough draft of a program and hope the community fills in the blanks for you.
Being open source isn't a bad thing and has certainly helped AMD. The reason why AMD is the popular choice for Linux is because AMD's driver work better than Nvidia's. The reason for this is because Valve has been a big contributor to AMD's drivers. AMD even recently open source ROCm, though too little too late if you ask me. But AMD does need to be a bigger contributor to their own software, and not depend on companies like Valve. A common theme with AMD is that they're afraid to spend serious money to develop Radeon graphics, and that's a problem. It's also unreasonable as Nvidia is one of the, if not the most valuable company right now because they've invested so much into graphics. AMD is missing out on the AI boom because they never invested enough resources into ROCm and other technologies. But AMD does need someone who can brain store new ideas, and not copy what Nvidia does and do it poorly.
 
Being open source isn't a bad thing and has certainly helped AMD. The reason why AMD is the popular choice for Linux is because AMD's driver work better than Nvidia's. The reason for this is because Valve has been a big contributor to AMD's drivers. AMD even recently open source ROCm, though too little too late if you ask me. But AMD does need to be a bigger contributor to their own software, and not depend on companies like Valve. A common theme with AMD is that they're afraid to spend serious money to develop Radeon graphics, and that's a problem. It's also unreasonable as Nvidia is one of the, if not the most valuable company right now because they've invested so much into graphics. AMD is missing out on the AI boom because they never invested enough resources into ROCm and other technologies. But AMD does need someone who can brain store new ideas, and not copy what Nvidia does and do it poorly.
But Valve shouldn't be the primary reason why AMD drivers work, Valve should be a contributor for sure they should be helping where they can but they shouldn't be doing the heavy lifting, that should be AMD's job.
My experience with AMD and open source has been "that's where AMD puts projects to die", they never provide them the resources they need to succeed and it is 99% carried by the community with AMD giving some basic oversight at best.
Then if one of those contributors manages to make it do something awesome, AMD leaps in with "Look at what our community has done, they are great we are great, opensource is great" Then AMD forgets about it for some time and those community members feeling burnt out move on and the project lingers again until somebody else tries to make it do something cool and the pattern repeats. But the software never progresses.

I can't say if AMD's Linux drivers work better, but I do know that I've never had a problem with the closed-source Nvidia drivers, they update frequently, and most issues have been solved by fixing some other program set not the drivers.

Regardless, AMD needs to relocate some resources to their software teams and put those teams to work, AMD is falling behind, and simply making a good accelerator is not enough right now.

But ROCm is a turd and I don't know how much they can reasonably polish that one.
 
Industries where working groups develop next gen API's work very well, and better stimulate competition in the marketplace.
So does working group making proprietary instead of open standard-API (say HDMI, it is a working group making a closed standard) or the smartphone industry, we cannot say it has not been vibrant, fast evolving with giant R&D budget despite a bunch of proprietary effort.

Does Tesla put that much money and effort in charging station if that give their cars absolutely zero advantage over the competition ? Always a balancing act, always compromise, never is ever obviously better in everyway. Should Tesla building a charging network been illegal, would making it proprietary to the point they just do not do it make people better for it ?
 
Who have you that idea? I grill 3-4 times a week all year round. I prefer grilling in the fall, spring or winter over summer time. I'd rather not sweat my balls off.

I neglected to consider different regions and climates. I don't think Ive ever grilled before June or after September.
 
I neglected to consider different regions and climates. I don't think Ive ever grilled before June or after September.
And this would not be for a grill, it is for a bbq, very long smoke (4-10-15 hours) where tracking the temps with alarm if something get too low or too high is interesting while doing something else (like sleeping or drinking in the pool.... your brisket hit 160 it's time to wrap it)
 
Last edited:
And this would not be for a grill, it is for a bbq, very long smoke (4-10-15 hours) where tracking the temps with alarm if something get too low or too high is interesting while doing something else (like sleeping)

My ovens and range have wifi that will notify you if someone turns them on, people think it's retarded but if you have kids you'll probably get why this is actually a pretty decent idea.
 
I'm so confused by the recent comments, is AMD taking on Weber next? We're going back to 28 nm!
 
I'm so confused by the recent comments, is AMD taking on Weber next? We're going back to 28 nm!
No. Tesla is going to acquire AMD. Will be announced any day now., and then build ARM-based chips to solve all the problems with self-driving cars.
 
But Valve shouldn't be the primary reason why AMD drivers work, Valve should be a contributor for sure they should be helping where they can but they shouldn't be doing the heavy lifting, that should be AMD's job.
I don't think Valve should need to be contributing at all. They contributed because they were going to use AMD for their Steam Deck. Valve were to the ones who helped make RADV, and also made ACO for AMD as well. I say this like Intel isn't also guilty of this as well.
My experience with AMD and open source has been "that's where AMD puts projects to die", they never provide them the resources they need to succeed and it is 99% carried by the community with AMD giving some basic oversight at best.
That is the case sometimes but not always. AMD did make AMDVLK, and is the primary contributor to it, but also hardly anyone uses it over RADV. They're also the main contributors to FSR and RadeonSi. ROCm wasn't even open source until a few months ago, and AMD isn't giving up on that. The issue I see is that AMD is treating these as side projects and not giving them the full attention they need.
Then if one of those contributors manages to make it do something awesome, AMD leaps in with "Look at what our community has done, they are great we are great, opensource is great" Then AMD forgets about it for some time and those community members feeling burnt out move on and the project lingers again until somebody else tries to make it do something cool and the pattern repeats. But the software never progresses.
I've never seen AMD gloat at anything, but then again who pays attention if they do?
I can't say if AMD's Linux drivers work better, but I do know that I've never had a problem with the closed-source Nvidia drivers, they update frequently, and most issues have been solved by fixing some other program set not the drivers.
There's a few cases I know of where Nvidia fails on Linux. One is Wayland, though I hear it's better now. Two is driver installation as it's never easy. I have screwed up Nvidia driver installs so often on Linux that I just don't bother installing them sometimes. Third is when a new game comes out and requires new changes to the drivers, which of course you can't because Nvidia is closed source. It's open source now, but realistically nobody is doing anything with it. Borderlands 3 was a nightmare for a lot of people to get running on Linux because something new had to be added to the drivers. Valve just quickly adds it to RADV and they're done, but Nvdia owners had to wait for Nvidia to do it which took a while.
nvidia linux driver install..jpg
Regardless, AMD needs to relocate some resources to their software teams and put those teams to work, AMD is falling behind, and simply making a good accelerator is not enough right now.
At this point if RDNA4 was better than the RTX-50 series in every way, it would still lose. Nvidia is cooking up something that will be exclusive to their hardware and make their cards better. Look at Ray-Tracing when Nvidia first introduced it we had Tom's Hardware that said you should get these cards for Ray-Tracing, even though you needed something like a 2070 or 2080 to get a playable frame rate. Eventually Nvidia creates DLSS to solve this problem and it's still solving Ray-Tracing performance for them to this day. It's now a reason to go buy an Nvidia GPU, just to have this feature, with or without using Ray-Tracing.
But ROCm is a turd and I don't know how much they can reasonably polish that one.
Their problem to deal with. They have money now and should invest it into these things. These are the same people who released the RX 7900 XTX for $1k but also sold the 7900 XT for $100 less, as if people are stupid enough to be tricking into that. This was such a massive screw up that they ended up making a 7900 GRE that they only sold to China, because they were hoping to milk America and Europe out of their money. They eventually released the 7900 GRE to the west, and still nobody wants to buy them. AMD is all out of good will from it's consumers. They either make a better product or Intel will. Who knows, maybe Jensen Huang's heart will grow three sizes larger and make their GPU's affordable.
 
Didn't AMD state they are hiring buttloads of new staff... specifically for the, um, software side? Iuno, I could be wrong...
 
Yeah, and I will win the Powerball lottery for like $45 billion, and tax-free.

I'm surprised that there are people who think Nvidia's going to make video cards...

Don't they still have like a 2-year supply of 3060s to move?
 
I'm surprised that there are people who think Nvidia's going to make video cards...

Don't they still have like a 2-year supply of 3060s to move?
No Nvidia already sold them, those unsold units are an AIB problem now. They ordered them, they produced them, it’s on them to sell them.
Jensen already told them off after the Crypto bust left the market flooded and made it clear he wasn’t bailing them out again.
 
I'm surprised that there are people who think Nvidia's going to make video cards...

Don't they still have like a 2-year supply of 3060s to move?
Am I reading this right? NVidia not going to bring out new models because they have all this old new stock still in the warehouse? They can always fire-sale it for like $50. :rolleyes:
 
  • Like
Reactions: Axman
like this
Am I reading this right? NVidia not going to bring out new models because they have all this old new stock still in the warehouse? They can always fire-sale it for like $50. :rolleyes:
and lose money. Nvidia is not very good at losing money. When you fire-sale tons of inventory at a loss, you bring down the perceived value (what customers expect to spend on what they want) of your product. Fire selling the 3060s at $50 makes the 4060 look a ton less compelling, then you'll have to fire-sell those, then people are used to spending sub-$100 for 4060s and then the 5060 comes out with pretty much the same performance at $799 and people won't want to spend that kind of money.

So Nvidia waits.
 
  • Like
Reactions: Axman
like this
Back
Top