You can grab a AMD eight-core Zen 3 CPU for as low as $139

[Admin]

Right now, the AMD Ryzen 7 5800XT CPU is available at Amazon for its lowest price—just $139 instead of the recommended $249.

You can grab a AMD eight-core Zen 3 CPU for as low as $139 : Read more

 

[rluker5]

Faster 12600kf is $143 and slower 12490f is $110 on Newegg right now. But both are in the same performance class.

 

[Shiznizzle]

Faster 12600kf is $143 and slower 12490f is $110 on Newegg right now. But both are in the same performance class.

IF you are looking at getting the AMD chip in this article then it is not likely you are even considering an Intel chip.

A vegetarian looks at the menu of a restaurant and gets bothered by a guy from across the street who has a steakhouse and is offering a steak.

 

[rluker5]

IF you are looking at getting the AMD chip in this article then it is not likely you are even considering an Intel chip.

A vegetarian looks at the menu of a restaurant and gets bothered by a guy from across the street who has a steakhouse and is offering a steak.

Of course somebody that already has an AM4 motherboard should get an AM4 chip, but most people are brand agnostic and getting new, so the comparison would be closer to going to the Thai place on one side of the street or the Hmong place on the other for them. Also for most the 12490 would give them a smoother, quieter and more luxurious experience for less money.

 

[Thunder64]

Of course somebody that already has an AM4 motherboard should get an AM4 chip, but most people are brand agnostic and getting new, so the comparison would be closer to going to the Thai place on one side of the street or the Hmong place on the other for them. Also for most the 12490 would give them a smoother, quieter and more luxurious experience for less money.

You sure aren't. You push Intel every chance you get. Tell me, how many upgrades were worthwhile from Alder Lake? Compare that to AM4. Intel plays socket roulette.

 

[rluker5]

You sure aren't. You push Intel every chance you get. Tell me, how many upgrades were worthwhile from Alder Lake? Compare that to AM4. Intel plays socket roulette.

That is nice how many CPUs AM4 went through, but maybe you are looking at it wrong.
I went from a Z97 with a 2015 5775c: and 2400c10 DDR3 https://www.anandtech.com/show/1619...ive-review-in-2020-is-edram-still-worth-it/25
straight to a Z690 with a 12700k and DDR5.
The Ryzen 3000 series would have been a sidegrade, everything before a downgrade, the 5000 series a definite upgrade, but I would have been upset if I just bought a 5800XT to want to replace it the same year with a 5800X3D.
I did do that with replacing my 12700k with a 13900kf and it really didn't save me money. What saved me money is not needing to upgrade for 6 years and a few chipsets. That is where I am now. I have no desire to go for the Z890 chipset, or upgrade to a 14900k. With the way games are going with the increased raytracing and frame gen:

it looks like my (run faster than this one) CPU will not need replacing until mid-late AM6. Saving me more money for GPUs. Although I could be tempted with something like 52 cores faster than mine just for hoarding reasons.

Isn't it better to not need an upgrade than to have more few hundred dollar ones available?

 

[Kindaian]

Intel plays core roulette as well. With a mix of "Perfomance" and "Economy" cores. So when you see a 12 core Intel chip, you need to check how many of those are performance ones, because the others are only good for low intensity tasks.

If you are looking for a Gamers CPU, that is mostly irrelevant, but if you are looking to run VMs on those cores, then the "Economy" cores are quite underpowered.

Would be nice to see performance stats with VMs running on all cores and then sum the total, instead of trying to run the stats on just one main OS with all cores available.

Load an hipervisor, and create the number of VMs as the CPU has cores, and then the number of VMs as the CPU has threads. And then on both scenarios, run the performance test suit. That would for sure give a more accurate picture of the raw power available.

And as a bonus point, add the actual power consumption of the system under load before, during and after the tests are run (and create a chart to compare the power used vs the performance of each of the cores / threads).

This would also allow comparing distinct architectures like ARM which also have the notion of big / small cores.