GPUs & CPUs & Enthusiast hardware: Questions, Discussion and fanboy slap-fights - Nvidia & AMD & Intel - Separe but Equal. Intel rides in the back of the bus.

⠠⠠⠅⠑⠋⠋⠁⠇⠎ ⠠⠠⠊⠎ ⠠⠠⠁ ⠠⠠⠋⠁⠛ said:
It would be better if it had five more PCIe slots, a couple PCI, and at least one ISA
Click to expand...
Yeah yeah.
This is genuinely a very cool piece of hardware. All you need to supplement to build a strong upper midrange productivity or gaming PC is some RAM, an SSD, a 90mm (?) fan for the heatsink, and a power supply (which can be a very minimal 30$ 300W FlexATX server thing from aliexpress or an ewaste reseller). You can upgrade it into a very capable gaming computer by just adding a graphics card, but the 780M it comes with is more than sufficient for "high" graphics on 1440p or "ultra" on 1080p (just get 6000+ MT/s memory so it won't be too bandwidth starved). My laptop has that same iGPU, and I regularly play games like BG3 on Ultra at its default resolution ("3k"), though with FSR set to performance (which produces no noticable artifacts on the 14" screen, but wouldn't be ideal on a desktop monitor).

If you want a retro PC with ISA and PCI, just buy a retro PC. No peripherals that plug into those slots work even half well with modern operating systems.
 
Sad day for my computer kiwi bros. My x570 based desktop that I've been using since late 2019 has confirmed memory errors. I found it out when large steam games > 8gb would fail to install and show a corrupt update file error. Memtest86 confirmed many failed addresses in my gskill 3600mhz sticks. Not too thrilled about its short demise.
 
  • Feels
Reactions: Bananadana, Judge Dredd, Wooo and 5 others
Fucking hell Intel.
i9.png
 
  • Horrifying
  • Lunacy
Reactions: The Backwards Man, ⠠⠠⠅⠑⠋⠋⠁⠇⠎ ⠠⠠⠊⠎ ⠠⠠⠁ ⠠⠠⠋⠁⠛, Wooo and 6 others
I'm stuck in the never ending loop of waiting for a newer/better waste of money for a gpu. Was considering the 4080Super but I'm not super sold plus its a fuckin grand

I hate computers
 
  • Feels
Reactions: Judge Dredd
The Ugly One said:
Again, I think it's just that they don't want to put off a launch any longer than they have to. You want to ensure your base design has no issues before you test a more demanding design.

I also don't know where you heard the technology wasn't ready at TSMC. 7003 EPYC was launched in March '21, 7003X EPYC in November '21, 5 months before 5800X3D in April '22. And months before the 7003X launch, AMD had shipped the entire first batch of Milan-X CPUs to Microsoft for installation in its HBV3 clusters. The reason 5800X3D wasn't until April '22 was the desktop division had to get in line behind the server division for this technology. At the time, they were selling every single 7003 they could make as fast as they could make it, and 7003X was just as hot. Plus the 3D V-Cache added a $1500-$2000 premium to the chip.

I believe 7800X3D is using 7nm L3 chiplets, so I think it's not directly competing with Genoa-X for resources there.
Click to expand...
What's the main difference between the 5000 and 7000 series? The 5000 by all accounts is still selling like crazy, so the tech clearly isn't that outdated yet
 
Car Won't Crank said:
Sad day for my computer kiwi bros. My x570 based desktop that I've been using since late 2019 has confirmed memory errors. I found it out when large steam games > 8gb would fail to install and show a corrupt update file error. Memtest86 confirmed many failed addresses in my gskill 3600mhz sticks. Not too thrilled about its short demise.
Click to expand...


In the earlier AM3 gen there was a few AMD bugs with memtest on certain AMD chipsets (think it was the later test sets like 5 and 6).

I got into the habit of testing RAM in Intel machines before tossing sticks or drawing conclusions. Or you could have bad ram + bad board. Either way its a rough one.
 
  • Informative
Reactions: Car Won't Crank
WelperHelper99 said:
What's the main difference between the 5000 and 7000 series? The 5000 by all accounts is still selling like crazy, so the tech clearly isn't that outdated yet
Click to expand...

5000 is a 7nm Zen 3 chip, 7000 is a 5nm Zen 4 chip. So 7000 series has more transistors per unit area and draws less power per GHz.

Probably An Autist said:
I'm stuck in the never ending loop of waiting for a newer/better waste of money for a gpu
Click to expand...

I did this for about seven years once.
 
  • Horrifying
  • Informative
Reactions: Judge Dredd and WelperHelper99
The Ugly One said:
5000 is a 7nm Zen 3 chip, 7000 is a 5nm Zen 4 chip. So 7000 series has more transistors per unit area and draws less power per GHz.
Click to expand...
Ah. General efficiency and effectiveness, but nothing truly drastic by the way you're talking. 7nm is still small by any standard.
 
WelperHelper99 said:
Ah. General efficiency and effectiveness, but nothing truly drastic by the way you're talking. 7nm is still small by any standard.
Click to expand...

It's all relative. TSMC N7 has a density of ~91.2 MTr/mm2. N5 has a density of ~108.2 MTr/mm2. So it is about an 18% difference. If you compare two chips at the same tier, like the 7950X and the 5950X, you can see the difference. It's meaningful, but it's not massive. I wouldn't expect a 5000 series Ryzen to be truly obsolete for a long time, perhaps not until 2030.

 
  • Like
  • Informative
  • Agree
Reactions: Bananadana, Post Reply, The Mass Shooter Ron Soye and 4 others
The Ugly One said:
It's all relative. TSMC N7 has a density of ~91.2 MTr/mm2. N5 has a density of ~108.2 MTr/mm2. So it is about an 18% difference. If you compare two chips at the same tier, like the 7950X and the 5950X, you can see the difference. It's meaningful, but it's not massive. I wouldn't expect a 5000 series Ryzen to be truly obsolete for a long time, perhaps not until 2030.

Click to expand...
I mean with how powerful CPU's are anymore, I hope it would still be viable for a while longer. Intel was stuck on 10nm forever, and those chips are still pretty solid.
 
The Ugly One said:
It's all relative. TSMC N7 has a density of ~91.2 MTr/mm2. N5 has a density of ~108.2 MTr/mm2. So it is about an 18% difference. If you compare two chips at the same tier, like the 7950X and the 5950X, you can see the difference. It's meaningful, but it's not massive. I wouldn't expect a 5000 series Ryzen to be truly obsolete for a long time, perhaps not until 2030.

Click to expand...
It also depends on what you're doing. Crunching numbers for crunching numbers sakes? Getting new processors is worth it. Building a gaming computer? Processor is very unimportant, the 3600 wasn't brilliant even when new but still runs almost all games just as well as the 7950X3D.
 
  • Agree
  • Informative
Reactions: WelperHelper99 and The Ugly One
snov said:
It also depends on what you're doing. Crunching numbers for crunching numbers sakes? Getting new processors is worth it. Building a gaming computer? Processor is very unimportant, the 3600 wasn't brilliant even when new but still runs almost all games just as well as the 7950X3D.
Click to expand...
That's always interesting to me. Games with 3d and all this shit is some how easy... but math? Actual calculations? That shit is hard and will make your computer chug even if it is modern if you're pushing it.
 
WelperHelper99 said:
That's always interesting to me. Games with 3d and all this shit is some how easy... but math? Actual calculations? That shit is hard and will make your computer chug even if it is modern if you're pushing it.
Click to expand...
@The Ugly One is who you want to discuss this with, he's very knowledgeable, but as far as I understand it it boils down to parallelisation. You can easily split "render a 3D scene" into multiple threads, and you can take shortcuts in the hardware since each pixel doesn't need to be perfect. That's not necessarily the case for math.
More and more simulations are taking advantage of GPUs, particularly the wide SIMD instructions, and in this case you have to get ridiculous server hardware to keep up with much cheaper and more accessible GPUs, but there are tasks that processors do more precisely.
 
  • Informative
Reactions: Brain Problems and WelperHelper99
snov said:
@The Ugly One is who you want to discuss this with, he's very knowledgeable, but as far as I understand it it boils down to parallelisation. You can easily split "render a 3D scene" into multiple threads, and you can take shortcuts in the hardware since each pixel doesn't need to be perfect. That's not necessarily the case for math.
More and more simulations are taking advantage of GPUs, particularly the wide SIMD instructions, and in this case you have to get ridiculous server hardware to keep up with much cheaper and more accessible GPUs, but there are tasks that processors do more precisely.
Click to expand...
I guess I can see how math is on a single thread. It would kind of be hard to split up what's basically a single task. A task you need to nail, every time. It can't be a little buggy like a game.
 
WelperHelper99 said:
I guess I can see how math is on a single thread. It would kind of be hard to split up what's basically a single task. A task you need to nail, every time. It can't be a little buggy like a game.
Click to expand...
Math isn't actually that single-threaded, at least not at the level our sims work at. Think something like simulating the flow of air through a turbine. If you represent a discrete volume of air as multiple smaller volumes and then step through the motion, you'll probably see what I mean. Each volume depends on its neighbours, but there are a lot of things you can do in parallel still. Engineering workstations still prioritise core count over single core FLOP.
 
  • Informative
Reactions: WelperHelper99
You must log in or register to reply here.
Back