Meaker

What you are hitting are the limits of the IO die I think. The 9800X3D showed some productivity gains as it was relieving some of the pressure on it.

That's like with cars saying you changed the oil and other accessories in half the time the manual states and are fine, that's still the failure of the manufacturer to give proper dates for service. It's a failure of the product, some of which were running in server environments with power limits and could still kill themselves. It's just not good enough IMO, same with the AMD side of things. Then to lie, take ages to admit something was going on and rejecting warranties is so scummy.

Intel way better, you get the 13th and 14th gen, with the 14th gen being the 13th gen just binned slightly higher and free corrosion in 13th gen amd degredation in both. A bargain.

You are better trying to push 6200/6400 with matching infinity fabric clocks (2066/2133) though IIRC?

https://www.3dmark.com/spy/47636486 My max was 1900-1950Mhz before voltage was the main limiting factor, the GPU core was at 60-65C.

It's usually a shunt resistor and monitor with an inbuilt limit on the readings on the BIOS. A hardware shunt mod would be a way around this usually.

Raid 5 for these 🙂 They are 7200RPM enterprise seagate drives. Pretty quiet actually, I booted everything up for a test and initialised an array and did some basic tests.

There are some interesting productivity cases where I think the I/O die is holding the 9000 series silicon back and the cache is helping ease that so that the 9950X3d which could actually really benefit given the restrictions. I'll be interested in the benchmarks.

I got 3x 12tb seagate enterprise drives and a controller card for my wife's upgrade partly because even with extra ports raid 5 is not supported.

My B650E-i was £240, in terms of chipset the B650E for ITX makes the most sense. Yes no TB4 but it's not something I use.

The 9950X3D with the improved cache die layering and 3d vcache on both CCDs could be the most interesting high end enthusiast chip for a while.

With the digital VRM bypass that DB was talking about there are some interesting tweaks to be made for high thread count workloads, benchmarking and overclocking. Oh also I have moved over to the B650E-I motherboard and am running PBO with a -25 curve offset now.

I mean it makes it cheaper to produce cutting out HT so all the reason for Intel, just not for the end user. A bit like some of the tech being tested here. I can see it all helping a lot in the future but it's an odd balance for now, but you have to do the switch over at some point.

I moved my 32GB kit into Alison's PC and picked up a 48GB kit of patriot 6000 CL36, it has lights so of course is 10% faster 😄

I just checked my old machine and turns out it was the CPU that died, fired right up with the 9600KF so I have a P870TM3 kit with original heatsinks and a 1080.