The computer and electronics super geek thread

[azrael]

Review embargo for Ultra7 270K Plus and Ultra5 250k Plus and the consensus appears to be "Why didn't you lead with THIS Intel?" Cheaper, much better in productivity and gaming (beating out AMD X3Ds in some scenarios) compared to the original 200-series Core Ultra release. Downsides are slightly higher power draws and it's coming out on the dead LGA1851 socket. Nova Lake CPUs are coming and this appears to be "Hey, we're still here!"-release. This release falls into the too-little-too-late bucket. Unfortunately, with all other components going up in costs (including CPUs), this release may not see much adoption. 

[David Hingtgen]

21 hours ago, JB0 said:

It has been freed from the constraints 

Unshackled AI, you say?

[davidwhangchoi]

ASRock motherboard still killing 9800X3D's 

 

 

[TangledThorns]

6 hours ago, davidwhangchoi said:

ASRock motherboard still killing 9800X3D's 

 

 

That is one way to kill your brand. I went with a MSI MAG 850M Mortar for my 9800X3D build last year and its been great.  

You can also save money at MSI's store with a CPU/mobo bundle too. Getting a 9800X3D for $399 is an outstanding deal! 

 

 

[davidwhangchoi]

7 hours ago, TangledThorns said:

That is one way to kill your brand. I went with a MSI MAG 850M Mortar for my 9800X3D build last year and its been great.  

You can also save money at MSI's store with a CPU/mobo bundle too. Getting a 9800X3D for $399 is an outstanding deal! 

 

 

that's a good deal for the combo.

[David Hingtgen]

I am almost purely MSI/Gigabyte/Corsair nowadays, when I have a choice.   (I would be 100% Gigabyte if I could, but they don't make everything, and I swear 2/3 of my last PC's have been "emergency" purchases so I have to take what I can get on short notice, not what I TRULY want)   

[mikeszekely]

On YouTube, ETA Prime pulled the graphene pad that's stock over the MacBook Neo's internals and cut the section that goes over the A18 out.  He then applied some Noctua thermal compound over the A18, then covered it with a strip of copper.  On the other side of the copper he used an Arctic TP-3 thermal pad to create a physical connection between the copper and the bottom of the Neo's aluminum chassis.  After doing so, he was able to get roughly 10-20% better scores on synthetic benchmarks.

For less synthetic results, he was playing No Man's Sky.  Totally stock, he was running the game on about medium settings and the Neo would quickly hit 105 degrees and start thermal throttling, resulting in around 30fps.  Just by replacing the stock pad with his copper-and-TP3 solution the A18 stayed in the mid-80s, which made throttling unnecessary and netting him nearly 60fps.

I don't have any sheets of copper just lying around, but I wonder how long it's gonna be before we start seeing professional kits that do this?  Seems like there's a lot of extra performance that can be wrung out of the A18 Pro just with better passive cooling.

[azrael]

1 hour ago, David Hingtgen said:

I am almost purely MSI/Gigabyte/Corsair nowadays, when I have a choice.   (I would be 100% Gigabyte if I could, but they don't make everything, and I swear 2/3 of my last PC's have been "emergency" purchases so I have to take what I can get on short notice, not what I TRULY want)   

ASrock’s lower end stuff doesn’t seem to cause as many problems like their high-end stuff does. 🤷‍♂️ And it’s the 9800X3D+870 motherboard that seems to be the most problematic for ASRock. 
 

40 minutes ago, mikeszekely said:

On YouTube, ETA Prime pulled the graphene pad that's stock over the MacBook Neo's internals and cut the section that goes over the A18 out.  He then applied some Noctua thermal compound over the A18, then covered it with a strip of copper.  On the other side of the copper he used an Arctic TP-3 thermal pad to create a physical connection between the copper and the bottom of the Neo's aluminum chassis.  After doing so, he was able to get roughly 10-20% better scores on synthetic benchmarks.

Yeah, he also attached an external liquid cooler. ZipTieTech and a few others just did the thermal pad mod (no additional cooling) and only got a <5% improvement on benchmarks. 

[mikeszekely]

1 hour ago, azrael said:

ASrock’s lower end stuff doesn’t seem to cause as many problems like their high-end stuff does. 🤷‍♂️ And it’s the 9800X3D+870 motherboard that seems to be the most problematic for ASRock. 
 

Yeah, he also attached an external liquid cooler. ZipTieTech and a few others just did the thermal pad mod (no additional cooling) and only got a <5% improvement on benchmarks. 

The numbers I quoted were just the thermal pad mod. When he added the cooler his synthetic benchmarks were closer to 17% better. The cooler also didn't make much difference while gaming; he had lower temps, but the CPU wasn't thermal throttling with just the thermal pad.

EDIT: Did ZTT only add a thermal pad? Maybe the copper ETA used is the difference.

Edited yesterday at 01:25 AM by mikeszekely

[azrael]

5 hours ago, mikeszekely said:

The numbers I quoted were just the thermal pad mod. When he added the cooler his synthetic benchmarks were closer to 17% better. The cooler also didn't make much difference while gaming; he had lower temps, but the CPU wasn't thermal throttling with just the thermal pad.

EDIT: Did ZTT only add a thermal pad? Maybe the copper ETA used is the difference.

The Geekbench score showed a ~9.7% uplift while doubling the FPS in No Man's Sky. After adding on the liquid cooler, Geekbench jumped up to 18.6% uplift. Cinebench multi-core gained 9.23% uplift with the copper mod and 19.08% uplift with the liquid cooler. Cinebench single core only showed ~5.78% with the copper-only and 23.51% uplift with liquid. The copper added more thermal mass so I can see some added uplift vs thermal pad-only but the end result is dumping the heat into the backplate. 

ZTT and the others just added a thermal pad to bridge the graphine sticker-pad to the chassis backplate. The copper ETA made a difference by adding more thermal mass. 

[davidwhangchoi]

so i just picked up Clair Exp 33 on pc and tried playing it with RTX DLSS Frame Gen, the input lag is annoying as hell.

Combat relies heavy on precise timing of parrying enemy attacks, I kept dying due to the input lag. I eventually had to become a clairvoyant parrying a second early before each enemy strike. 

 

Decent Deal that's currently in stock as of now:

STORMCRAFT - Gaming PC PHANTOM INTEL Core Ultra 7 265F NVIDIA RTX 5080 2TB SSD 32GB 6400MHz RGB 360mm AIO Windows 11 Home - Black $2499.99

Edited 11 hours ago by davidwhangchoi

[mikeszekely]

In my experience, using DLSS to render at a lower resolution and then upscale to your monitor's resolution works pretty well now. But I avoid frame gen if at all possible.

[davidwhangchoi]

6 hours ago, mikeszekely said:

In my experience, using DLSS to render at a lower resolution and then upscale to your monitor's resolution works pretty well now. But I avoid frame gen if at all possible.

i am using both but am going to turn off frame gen as you suggested.