News AMD Ryzen 7 9800X3D beats the 7800X3D by 26% in leaked Blender benchmarks — outpaces even the current-generation Ryzen 7 9700X by 11%

[Admin]

Improved thermals and higher clocks thanks to AMD's 2nd Generation V-Cache allow the Ryzen 7 9800X3D to pull ahead of its predecessor by almost 26% in Blender.

AMD Ryzen 7 9800X3D beats the 7800X3D by 26% in leaked Blender benchmarks — outpaces even the current-generation Ryzen 7 9700X by 11% : Read more

 

[Notton]

Even if real, if you buy the 9800X3D to run blender, you are doing it wrong.
If you're going to run productivity or mixed, get the i7.

 

[nightbird321]

Even if real, if you buy the 9800X3D to run blender, you are doing it wrong.
If you're going to run productivity or mixed, get the i7.

Or the 9950x3d for the best of both worlds if you use core parking to fix windows issues. This article casually leaves out the 550-590 score of the 7950x/x3d.

 

[AngelusF]

If you're using any sort of CPU to run Blender you're doing it wrong.

 

[TheHerald]

Last years i7 is 30% faster than the yet unreleased R7.

 

[Jame5]

The memory support sees an upgrade to DDR5-5600 (JEDEC) at one DIMM per channel. Using two DIMMs per channel, however, will drop your speeds to DDR5-3600 levels.

Serious question: When did consumer platforms start rating memory speeds like server platforms? It used to be whatever RAM speeds the platform was rated for it would work regardless of DIMM population at JEDEC speeds.

Now we are getting massive performance cliffs if you need more than 2 DIMMs (1 DPC) in a system. Is the desire for higher clocks somehow breaking the ability for IMCs to run 2DPC at standard JEDEC speeds?

 

[KraakBal]

"However, the 9800X3D easily outclasses the Ryzen 7 9700X this time by almost 11%. What's more bizarre is that the 9700X has a higher boost clock at 5.5 GHz."

No, the 9700x does NOT run at 5.5 for all core max-out workload, but just a single core. It would be less than the 9800x3d due to differences in tdp. Would think tech journalists would know this by now

 

[KraakBal]

Last years i7 is 30% faster than the yet unreleased R7.

Assuming this is not bait, and you are just ignorant: you are comparing 2 cpus with very different cores counts. Disable those e-cores or compare with 9900/9950x or their x3d variants to be fair, in this kind of workload

 

[TheHerald]

Assuming this is not bait, and you are just ignorant: you are comparing 2 cpus with very different cores counts. Disable those e-cores or compare with 9900/9950x or their x3d variants to be fair, in this kind of workload

Ah, the 9800x 3d has less cores? I assumed since it's going to be 30% more expensive than the i7 it should have a similar core count. My bad

 

[tamalero]

Or the 9950x3d for the best of both worlds if you use core parking to fix windows issues. This article casually leaves out the 550-590 score of the 7950x/x3d.

That bias is still there lmao.

 

[YSCCC]

Assuming this is not bait, and you are just ignorant: you are comparing 2 cpus with very different cores counts. Disable those e-cores or compare with 9900/9950x or their x3d variants to be fair, in this kind of workload

Fans can never be reasoned. But I am wondering with the uplift compared to last gen X3D being quite good, how the 9900X3D and 9950X3D will fare. Though price wise they are likely to skyrocket as ARL and RPL refresh isn't competitive

 

[KraakBal]

Ah, the 9800x 3d has less cores? I assumed since it's going to be 30% more expensive than the i7 it should have a similar core count. My bad

Just like a good strong v8 engine that might be more expensive than 20-30 donkeys that can do more work on a larger area in the same time?

 

[helper800]

Last years i7 is 30% faster than the yet unreleased R7.

25.3% was actually the metric. The 14700k may have more MT performance for productivity, but considering the 9800X3D may be about 25.3% faster in games you have to look at the use case for the CPUs before the price to see if it's worth the cost.

 

[TheHerald]

25.3% was actually the metric. The 14700k may have more MT performance for productivity, but considering it may be about 25.3% slower in games you have to look at the use case for the CPUs before the price to see if it's worth the cost.

25% slower, which doesn't mean 25% faster.

 

[TheHerald]

Just like a good strong v8 engine that might be more expensive than 20-30 donkeys that can do more work on a larger area in the same time?

If the donkeys can do more work I'll take the donkeys. Not even a question about it.

 

[helper800]

25% slower, which doesn't mean 25% faster.

Read it again, I used more concise language for you.

 

[KraakBal]

If the donkeys can do more work I'll take the donkeys. Not even a question about it.

Different usecases. A fast car will get you to work in the morning on time, allow you to roadtrip the country in a reasonable time.

Unless you plan to do video transcoding, compression/decompression work or run vms for your customers all the time, and 8 modern cores is still too slow to make you enough money on that, the other 99% of compute tasks you might want to run will benefit from less but faster cores/threads.

 

[Notton]

Or the 9950x3d for the best of both worlds if you use core parking to fix windows issues. This article casually leaves out the 550-590 score of the 7950x/x3d.

7950X3D is in a totally different price class and should be compared to the U9 285K and i9-14900K/S

 

[TheHerald]

Read it again, I used more concise language for you.

Your math is off. The article says that the 9800x3d is 25. 3% slower, which makes the i7 35% faster. I was generous and said 30.

 

[helper800]

Your math is off. The article says that the 9800x3d is 25. 3% slower, which makes the i7 35% faster. I was generous and said 30.

You are correct actually. I forgot they were using baselines... Dohhh!

 

[jeremyj_83]

Serious question: When did consumer platforms start rating memory speeds like server platforms? It used to be whatever RAM speeds the platform was rated for it would work regardless of DIMM population at JEDEC speeds.

Now we are getting massive performance cliffs if you need more than 2 DIMMs (1 DPC) in a system. Is the desire for higher clocks somehow breaking the ability for IMCs to run 2DPC at standard JEDEC speeds?

This has been the normal many years. Motherboards for Sky Lake on the H110 series were reducing speeds from 2133 > 1733 when 2DPC were used. Sky Lake was launched back in 2015.

 

[Manumarblanc]

https://www.techpowerup.com/review/intel-core-ultra-9-285k/32.html

https://www.cpubenchmark.net/high_end_cpus.html

https://www.techpowerup.com/forums/attachments/1729784946332-png.368752/

 

[Jame5]

This has been the normal many years. Motherboards for Sky Lake on the H110 series were reducing speeds from 2133 > 1733 when 2DPC were used. Sky Lake was launched back in 2015.

I guess I'm just confused since my 5900x runs 4x 32GB DDR4 DIMMS at stock DDR4-3200 speeds without any tweaks or adjustments.

 

[jeremyj_83]

I guess I'm just confused since my 5900x runs 4x 32GB DDR4 DIMMS at stock DDR4-3200 speeds without any tweaks or adjustments.

You got lucky on the silicon lottery then. Most of the time running 4 DIMMs with Ryzen will drop your speeds to 2666 or something similar.

 

[hannibal]

All in all this is as earlies leaks predict. Biggest upgrades are in multicore productivity and smallest in gaming compared to 7800x3d.
Rather promising in any way!