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

Just looking into GN's numbers a bit. I think he is right that the efficiency gains here are pretty underwhelming. One reason is there supposedly is only about a 6% transistor density improvement going from TSMC N5 to N4P. (This is what people are citing: https://www.techinsights.com/products/ace-2204-801). TSMC claims a 22% improvement in power efficiency.

But he is also directly comparing unlike processors. The higher you clock a CPU, the less efficient is, so it is not surprising at all that the 9950X, with its clock of 4.3-5.7 GHz, is less efficient than the lower-end CPUs with slower clock speeds. he doesn't mention that in his 7-Zip section. Notably, the 9700X (3.8-5.5 Ghz) is quite a bit more efficient than the 7700X (3.8-5.3 GHz), about 38% more efficient. That's not nothing.
The easiest way to compare the efficiency of Zen 4/5 is to compare the 65W Ryzen 7 7700 to the 65W Ryzen 7 9700X, although if the default TDP gets raised to 105W by AGESA 1.2.0.1A, suddenly the 9700X is a true 7700X competitor. :P They did compare to 7700 non-X in the video I linked, as well as turn on PBO for the 9700X which does whatever that does.

Despite TSMC N4P allegedly being only "6%" more dense than TSMC N5, the Zen 5 CCD is said to have about +28% density. More like +28% transistors, +28.7% density from a slight die size decrease:

CodenameCoresDie SizeTransistor CountNodeTransistor Density
Ryzen 7000 'Durango'8 Zen 471 mm^26.5 billion5 nm92.9 MTr/mm^2
Ryzen 9000 'Eldora'8 Zen 570.6 mm^28.315 billionN4P117.78 MTr/mm^2

How? The "6%" from TSMC is probably an underestimate due to averaging logic, SRAM, and analog transistors. And maybe they're using a different cell library or something. L2/L3 cache amounts in the CCD haven't changed. I think they need to be doing something to decrease empty space in there...

Here's one discrepancy. Tom's Hardware is sourcing from a German tech outlet, and they're saying that the node used could be TSMC N4X.
 
Last edited:
Might be better to just say 'process' at this point instead of node, also I feel like TSMC has not been moving the needle much lately in general
 
  • Like
Reactions: Fcret and Vecr
Hard to believe but Secure Boot BIOS security has been compromised on hundreds of PC models from big brands because firmware engineers used four-letter passwords
Ars Technica reports that, "researchers from security firm Binarly revealed that Secure Boot is completely compromised on more than 200 device models sold by Acer, Dell, Gigabyte, HP, Intel, Lenovo, Supermicro and others. The cause: a cryptographic key underpinning Secure Boot on those models that was compromised in 2022." Ouch.
Part of the problem, as we understand it, is device makers basically using the same old keys over and over again. To quote Binarly, the security failure involves, "no rotation of the platform security cryptographic keys per product line. For example, the same cryptographic keys were confirmed on client and server-related products. Similar behavior was detected with Intel Boot Guard reference code key leakage. The same OEM used the same platform security-related cryptographic keys for firmware produced for different device manufactures. Similar behavior was detected with Intel Boot Guard reference code key leakage."

The report includes a list of hundreds of machines from the brands mentioned above that have all been compromised by the leak. For the record, some of those systems include Alienware gaming desktops and laptops. Security experts say that for those devices that use the compromised key, it represents an unlimited Secure Boot bypass allowing malware to be executed during system boot. Only a direct firmware update for each device can secured affected devices.
 
Last edited:
I have a question for the more tehnical people:

So, while trying to find ways to stop my laptop Shidtel CPU from croaking, I stumbled upon several posts on platforms by a user named Da-BOSS.
Reddit acount: https://www.reddit.com/user/THEBOSS619/

He claims to be a ex kernel dev for a phone. Apparently he has made a tool that updates the microcode, something which he seems not to open to disclose on the ROG forum, since he larps as a distributor there.


Now I found this to be odd, as we were told the microcode would be featured in a BIOS update, and those clearly don't come in the form of a russki made .bat script.

But, looking through the forum, it genuinely seems like a bunch of (likely malicious) horseshit.

FIRST OF ALL, the update clearly does nothing to lower the voltage, as one retard who ran a fucking .bat he got from a stranger found out:

7.png


2.png


AND SECONDLY, the guy is himself sketchy as fuck. He shills for lenovo like his shithole of a country depends on it. And he is clearly a russian, given that his account is set to use the russian language, which you can see if you hover on the elite icon next to his name.

Introduction aside,

My question to the technical bros is:

Is this cheeki-breeki handing out crypto miners or something disguised as a update or what?
Either way I am not going to download this shit as it's sketchy, but I would like to hear what you think.
 
I have a question for the more tehnical people:

So, while trying to find ways to stop my laptop Shidtel CPU from croaking, I stumbled upon several posts on platforms by a user named Da-BOSS.
Reddit acount: https://www.reddit.com/user/THEBOSS619/

He claims to be a ex kernel dev for a phone. Apparently he has made a tool that updates the microcode, something which he seems not to open to disclose on the ROG forum, since he larps as a distributor there.


Now I found this to be odd, as we were told the microcode would be featured in a BIOS update, and those clearly don't come in the form of a russki made .bat script.

But, looking through the forum, it genuinely seems like a bunch of (likely malicious) horseshit.

FIRST OF ALL, the update clearly does nothing to lower the voltage, as one retard who ran a fucking .bat he got from a stranger found out:

View attachment 6312187

View attachment 6312188

AND SECONDLY, the guy is himself sketchy as fuck. He shills for lenovo like his shithole of a country depends on it. And he is clearly a russian, given that his account is set to use the russian language, which you can see if you hover on the elite icon next to his name.

Introduction aside,

My question to the technical bros is:

Is this cheeki-breeki handing out crypto miners or something disguised as a update or what?
Either way I am not going to download this shit as it's sketchy, but I would like to hear what you think.
It's a retarded idea to trust any custom update, regardless if it works or not. And it wouldn't surprise me if it works, because those voltage results mirror what people are seeing on desktops; Intel didn't fix the problem, Intel are just trying to mitigate the problem in the short term so they can sweep it under the rug.
 
The easiest way to compare the efficiency of Zen 4/5 is to compare the 65W Ryzen 7 7700 to the 65W Ryzen 7 9700X

You have to be careful here. Clock management can be very different across models and generations - I know I've been caught off guard by processors with barely-changed max clock speeds being significantly faster due to just how long they stayed at higher clocks. I've seen myself a BIOS change that changed the thermal limits of a chip such that it ran about 2x hotter and 40% faster - boost and base clocks were the same, base TDP was the same. So you need to look at both how fast the task gets done and how much energy is consumed. And taking all this into account, we do see, via TH, the Ryzen 9700X beating the 7700 and 7700X both.

1723779515673.png

1723779637165.png

It would be nice to see a mix of single-threaded, multi-threaded, compute-bound, and memory-bound benchmarks here.

as well as turn on PBO for the 9700X which does whatever that does.

PBO changes how the CPU manages power so that it will run at higher clock speeds for longer periods of time. Here's a very good illustration of the difference. PBO doesn't change clock speeds, just how they're managed.

1723778107227.png

sauce: https://www.pugetsystems.com/labs/a...ermals-and-content-creation-performance-2373/
 
It's time to accept that Zen 5 is broken... on Windows at least.



The bug described above also affects Zen 4 and doesn't change positioning of the two beyond 1%.
This reminds me of how the Windows scheduler needed a patch to run AMD FX processors properly on Windows 7.

Older Ryzen CPUs could also be affected by this bug. According to this comment showing 5800x3d (Zen 3) benchmarks:
Cyberpunk 2077 built-in benchmark:

Stock: 1%low - 93.3 FPS / Average - 179.9 FPS / Max - 246.4 FPS
Admin: 1%low - 96.3 FPS / Average - 195.4 FPS / Max - 271.6 FPS

Ratchet & Clank:

Stock: 1%low - 119.4 FPS / Average - 214.4 FPS / Max - 295.7 FPS
Admin: 1%low - 115.9 FPS / Average - 235.8 FPS / Max - 310.1 FPS

Assetto Corsa Competizzione:

Stock: 1%low - 112.2 FPS / Average - 143.6 FPS / Max - 214.2 FPS
Admin: 1%low - 120.9 FPS / Average - 154.5 FPS / Max - 223.1 FPS

LOTF2:

Stock: 1%low - 92.1 FPS / Average - 143.1 FPS / Max - 192.6 FPS
Admin: 1%low - 92.3 FPS / Average - 149.4 FPS / Max - 206.5 FPS

Ghost of Tsushima:

Stock: 1%low - 106.3 FPS / Average - 127.1 FPS / Max - 156.4 FPS
Admin: 1%low - 107.8 FPS / Average - 128.7 FPS / Max - 156.2 FPS
Now I am wondering if this is a Windows Admin overhead problem. It would be interesting to see if there are any Intel CPUs that show a significant framerate increase when running games on the Admin account.
 
I'm genuinely surprised, those are some pretty decent differences, with over 20 fps max on the high end.

Do we know if it's just a game thing, or is it a intense workload thing, like Blender or something?
 
Is there a big difference between game engines and workload software in terms of code quality and processing efficiency?
 
It's time to accept that Zen 5 is broken... on Windows at least.




The bug described above also affects Zen 4 and doesn't change positioning of the two beyond 1%.

Actual fucking lmao. I remember testing my PC with Win10 LTSC and then switching to Linux Mint Cinnamon (22). The performance difference is quite noticeable, even when not in gaming where you can't even set your screen settings.

Anyway, for a friend: anyone who can recommend a Ryzen laptop? I see large swaths of intel laptops, my friend wants to play games normally (like GTA V) but I might end up buying a shit laptop from yesteryear or whatever so I'd like to hear some suggestions. Friend isn't in an electricity-confident place (i.e prone to blackouts) so no desktops for now.
 
Last edited:
Anyway, for a friend: anyone who can recommend a Ryzen laptop? I see large swaths of intel laptops, my friend wants to play games normally (like GTA V) but I might end up buying a shit laptop from yesteryear or whatever so I'd like to hear some suggestions. Friend isn't in an electricity-confident place (i.e prone to blackouts) so no desktops for now.

My experience with my Ryzen 6800U laptop (which I still have) is that it sucks down power like no tomorrow and runs blazing hot, even when just browsing the web, because of awful clock management. Plenty powerful, though.
 
My experience with my Ryzen 6800U laptop (which I still have) is that it sucks down power like no tomorrow and runs blazing hot, even when just browsing the web, because of awful clock management. Plenty powerful, though.
Anyway, for a friend: anyone who can recommend a Ryzen laptop?
I have a 5800H in a Lenovo Legion (16ACH6H) laptop and that has thermal issues too. The worst is the dedicated GPU's cooling is shared with the CPU, so even if the CPU is lightly loaded, it gets roasted up to 100C when the GPU is loaded doing ML tasks.

It's otherwise a good laptop. 2560x1600 165 Hz screen, can support 2 NVMe SSDs and has user replaceable SSDs, Wi-Fi card and memory. It's not super portable though, heavy laptop and power brick due to the power requirements as it has a 3070 GPU.

Has been reliable for me, did a lot of traveling with it no issues and Linux compatibility is mostly fine except for the Wi-Fi card (I replaced it with an Intel one because of that). There were issues with not being able to control the backlight but that has fixed itself and if you're a Winfag you won't have to deal with that shit anyway. I never use the Ryzen APU, all dedicated GPU as apparently GPU switching is janky in Linux and I always use it plugged into AC anyway.
 
Last edited:
  • Lunacy
Reactions: Susanna
I got a question. I built my first computer and I’m having an issue, maybe this isn’t the place for it but whatever I’ll try anyway.

I have this GPU, it’s a Asus RTX 3060, I bought it used from a friend, I know he wasn’t using it to mine and this thing is hot as fuck under full load. Like 90, easy. I’ve undervolted it, I’ve repasted it, my cooling situation is fine, I have no idea why it’s this goddamn hot and frankly it scares me a bit that I’m going to fry my expensive thing.

Am I just bugging out over nothing or am I dealing with a situation?
Old, I know, but you can activate a custom fan curve when under heavy load to bring the temps down at the price of noise. It's just one click in Afterburner or whatever it's called. Oh, and is it a single, dual or triple fan card?
 
  • Like
Reactions: Brain Problems
Back