AMD Ryzen 9 5900X + Ryzen 9 5950X Dominate On Linux

[f0rmat]

I am now confused if concurrency is easier or more difficult in Rust ...

Depends on if concurreny is done in Rust or if there is Rust on the concurrency since concurrency depends on the "bare metal."

[bridgman]

1.) What about the pre-Bulldozer days around 2007 where AMD had enjoyed their last heydays? They invested into the wrong architecture and made other wrong strategic and technical decisions which led to near bankruptcy later on. According to your logic they could never fail because they earned more money in that time frame and would therfore improve their standing continously. To the contrary, they could come up with the Ryzen architecture while being low on cash and with a lot less R&D budget than Intel. Sometimes it is up to focus and clever engineering to make something better than your competitior. They proved that they can achieve that with small teams. There is obviously a lot of room to grow, still. Their GPU division and software stack is one area where they could profit from more investments and more hands on the job (but maybe better organization and better work-flows would also yield better results, it is all about efficiency there, too?!).

With respect, we were already well into cash shortages and layoffs by 2007. The K10 architecture was just launching in 2007 and we lost ramp-up time via the TLB HW bug. AMD was deep into survival mode by the time Bulldozer development started.

If you want to say that Phenom and Barcelona were crap and they were developed when AMD was doing well that would be fair, but I don't hear that said much.

3.) You overestimate the developement costs for the Zen 3 chiplet as their main effort would have been the IP design and transition to the 7nm process which they had to do for Zen 2 anyway and could derive from that work later on. Of course the development of that new chiplet costs money and they had to order new masks for the Zen 3 chiplet. But the design process is comparatively cheap when compared to the overall manufacturing costs. And these are largely determined by the costs per wafer from TSMC which should be well below where they were last year when they launched Zen 2. They are now enjoying the benefits from their chiplet approach to keep their costs down and amortize large parts of their investments into Zen 2 over two generations, which is clever from their business perspective but we customers don't profit from these cost savings and have to pay a higher entry price. We also don't know for sure if and when a 5600 for 200 EUR arrives at the market. That is definetly the one to get if gaming is the only concern.

It might be worth reading up more about Zen3 architecture - the CPU core was a bottom-to-top revision with non-trivial development costs. Saying that "the design process is comparatively cheap when compared to overall manufacturing costs" is closer to wishful thinking than fact.

[L_A_G]

If you can clearly relate it to the PSU burning out it can be some faulty part on the motherboard. Might be a malfunctioning powersupply IC which only flicks off once in a while and takes down CPU or RAM. It is almost impossible to find it. This will cause Kernel Panic or BSOD without any warning.

A power-related IC like a MOSFET failing would fit the bill as the board is an MSI X370 Carbon that got a bunch of flack from people who review boards on their choice of ICs and saw that MSI cheap'd out on them when designing this particular board.

(But to be sure use the CPU and RAM on another Board. If it persists than you can play the swap game again with CPU or RAM. Eitherway it takes a while and you need new or borrowed components.)

That's the problem, I don't have any ones going spare nor know someone who could borrow me an AM4 board. Even my old machine, which I still have the leftovers of, isn't any help as it was a Nehalem-based hodgepodge of new and second hand components I built when I was still at university. It won't even accept the RAM.

May I ask what PSU it was? If it was a cheap one/no name/very old one...never buy cheap ones otherwise they might take down your whole system.

Admittedly it was a somewhat cheap one, Corsair VS650, but at least it wasn't a no-name brand and I learned my lesson there (replaced it with a much more expensive Fractal Design Ion+). Being a carry-over from my previous 2014 build it's obviously long past the 3 year warranty stated on Corsair's website. My focus trying to fix it this weekend is the HDD as the kernel panics are fairly similar to when the HDD on my old Macbook Pro died on me way back when. Everything freezes and stops working except the cursor, which I can still move around the screen. Better yet, unlike the PSU, CPU, GPU and motherboard it's still well under warranty as I only replaced it 2 years ago and it has a 5 year warranty.

[f0rmat]

If you can clearly relate it to the PSU burning out it can be some faulty part on the motherboard. Might be a malfunctioning powersupply IC which only flicks off once in a while and takes down CPU or RAM. It is almost impossible to find it. This will cause Kernel Panic or BSOD without any warning.
(But to be sure use the CPU and RAM on another Board. If it persists than you can play the swap game again with CPU or RAM. Eitherway it takes a while and you need new or borrowed components.)

Without PSU failure I have had a smiliar behaviour with my old AMD FX 8350 + MSI 990FX-GD60 board. plus reseting the board behaved strange.
My guts told me - Motherboard. But my first test was RAM - not the issue ...then new PSU (Coolermaster already 10years+/ Replaced by a new SuperFlower). Not the issue. .. considering the age I decided to buy a new Mobo+CPU (and RAM)

May I ask what PSU it was? If it was a cheap one/no name/very old one...never buy cheap ones otherwise they might take down your whole system.

I used to have occasional issues with motherboards when a power supply blew. I spent a f**kload of time researching about the capacitors used in the power/rectifier bridges. Then I went to UPS systems on my desktop/home server. Never blew another PSU again - and before I used to blow one about every 6 months - and they were not cheap ones. (I had myself, my wife, and four kids all with computers and two home servers). It is amazing how flaky power is with its dirty sine waves. After I went to using highly rated UPSs on my systems, I never blew a power supply again - and some of those PSUs (like Thermaltake and stock SuperMicro) ran for over eight years without an issue. The other advantage is that I plugged my cable box and modem and router in them so when the power "flickered" I did not lose network conductivity or have to wait 20 minutes for the cable box to reboot.

[Raka555]

It was tested against a $999 Intel® Pentium® Processor Extreme Edition 955

https://ark.intel.com/content/www/us...6-mhz-fsb.html

Some more facts about how the "innocent" AMD behaved when they had the upper hand ...


"The North remembers ..."

[f0rmat]

With respect, we were already well into cash shortages and layoffs by 2007. The K10 architecture was just launching in 2007 and we lost ramp-up time via the TLB HW bug. AMD was deep into survival mode by the time Bulldozer development started.

If you want to say that Phenom and Barcelona were crap and they were developed when AMD was doing well that would be fair, but I don't hear that said much.



It might be worth reading up more about Zen3 architecture - the CPU core was a bottom-to-top revision with non-trivial development costs. Saying that "the design process is comparatively cheap when compared to overall manufacturing costs" is closer to wishful thinking than fact.

I am glad to see AMD doing well - I remember all of the discussion at that time of AMD entering and trying to control the "budget" CPU market because of those very issues you mentioned. And indeed, I used many systems with AMD APU just shortly after that time that "just worked." My old HP laptop from 2011 or 2012 has an APU 10-4800 that has NEVER let me down. Ups and downs are business. AMD Athlon architecture put competition back into an industry that Intel, up to that point, had owned. If I remember correctly, AMD got on the map with the K6 architecture which were Intel clones under license from Intel (maybe it was earlier architecture). I even looked at a K6-2 build until the K7 Athlon came out. (I spent 400 USD on a K7 700MHz for that build - I thought the wife was going to kill me ). But I do remember the "WTF" from Intel people and computer enthusiasts when AMD, not Intel, broke the GHz barrier and again when AMD moved the front side bus to the CPU die eliminating the need for the north bridge. Made computer building much easier - no more matching north bridge to CPU to everything else. Am I a "die hard" fan boy, no. Am I believer in competition, hell yes. Competition brings out the best in a company and ultimately provides what is best for the consumer.

On another note, I remember seeing an April's Fool Day article on some tech site (maybe the old site 2CPU.com - I do not recall) that claimed (tongue in cheek) that ATI bought AMD. I and my friends thought it was hilarious. Not soon after, AMD announced that it had bought ATI. I and my friends thought that the irony of that was also hilarious.

[bridgman]

Some more facts about how the "innocent" AMD behaved when they had the upper hand ...

Yep... as the article you quoted clearly says, we launched high end products first and then followed them up with less expensive and broadly affordable products.

This has been standard practice for most product launches in the last decade - start with high end SKUs then define lower end parts once you have shipped enough high end parts to have a handle on what the yields will be at lower speeds & functionality.

[pmorph]

I don't get these complaints about pricing. Now Intel is again forced to reduce their pricing, and AMD will respond appropriately.
That's how it SHOULD be.
NOT so, that one party drops the prices just as a nice gesture.
NOT so, that one party keeps the prices high, just because there is no real competition.

I guess Intel dominance was so long that people forgot how free markets are supposed to work.

[f0rmat]

I don't get these complaints about pricing. Now Intel is again forced to reduce their pricing, and AMD will respond appropriately.
That's how it SHOULD be.
NOT so, that one party drops the prices just as a nice gesture.
NOT so, that one party keeps the prices high, just because there is no real competition.

I guess Intel dominance was so long that people forgot how free markets are supposed to work.

I love the free market economy and competition. You are on the money.

[Raka555]

Yep... as the article you quoted clearly says, we launched high end products first and then followed them up with less expensive and broadly affordable products.

This has been standard practice for most product launches in the last decade - start with high end SKUs then define lower end parts once you have shipped enough high end parts to have a handle on what the yields will be at lower speeds & functionality.

I do not favour Intel or NVdia or AMD over the other.

I know It is all about business.

Any of these companies will behave the same when facing the same circumstances.
AMD had to be the "value brand" for the past 15 years to survive.
Before that they were the "Premium product" and did not give an inch and Intel had to play the "value brand" role.
People just remember Intel now as the bad guy and AMD as the "saviour".
They don't realise that the roles were once reversed and that AMD will go into "Premium product" mode again when it suits them.