CPU RYZEN PRO 4750G/4650G
- Thread starter REVOCCASES
- Start date
i think it's also possible that renoir will get a very limited diy launch and we will go straight to cezanne (zen3 apu) in 2020/early 2021.
a0 stepping is ready, amd wants to unify the naming, some motherboard makers already said cezanne will be supported on 500 series chipsets. to me this seems like the next logical step.
a0 stepping is ready, amd wants to unify the naming, some motherboard makers already said cezanne will be supported on 500 series chipsets. to me this seems like the next logical step.
I think it's worth to wait for the next APU generation (graphics performance wise).
Edit: as I found out, my 3200G graphics performance is so poor because the AsRock A320TM-ITX I used only supports single channel RAM. The JUBO A300 I used for testing the 4650G seems to support dual channel.
Edit: as I found out, my 3200G graphics performance is so poor because the AsRock A320TM-ITX I used only supports single channel RAM. The JUBO A300 I used for testing the 4650G seems to support dual channel.
Last edited:
Any thoughts why my i5-1035G can beat the 3200G graphics performance? Using DDR4 3200 on all systems.
Edit: did another few real game benchmarks. The 4650G's Radeon is only 20 to 30 percent faster compared to the i5 which is only a mobile processor and locked to 15W TDP.
I've honestly expected more of Renoir.
Edit: did another few real game benchmarks. The 4650G's Radeon is only 20 to 30 percent faster compared to the i5 which is only a mobile processor and locked to 15W TDP.
I've honestly expected more of Renoir.
Last edited:
Tiger Lake beats Renoir:
AMD, it's time to improve your APUs graphics...
Golem.de: IT-News für Profis
www.golem.de
AMD, it's time to improve your APUs graphics...
Heaven is an older benchmark but legit for comparison from my POV. Especially since I was planning to run some older games with my APUs.
Not an apples to apples comparison as Linux, OpenGL, and 2400G. Seems like tests I've seen on Phoronix site the Windows/Linux scores were close at the time for 2400G IIRC with tests on Toms HW.
Any idea why my results are much worse than yours? All fresh Windows installs with latest drivers.
Edit: difference seems you did not enable 8xAA. Without AA I also get much better results... but still, the 15W i5-1035G beats the 65W 3200G
Last edited:
Yeah that cut me in half pretty much, inline with your results. I have an 8500T HP Prodesk might try this on, but I doubt it'll do half this.
Yeah that cut me in half pretty much, inline with your results. I have an 8500T HP Prodesk might try this on, but I doubt it'll do half this.
How is your RAM configuration on this system?
I wouldn't expect too much from the 8500T but with recent CPUs Intel improved their iGPU performance a lot.
Edit: my poor results for 3200G is related to the A320TM-ITX board which only runs on single channel RAM despite AsRock saying it would support dual channel:
Motherboard - AM4 Thin ITX motherboards, They do exist!
I can't find any US retailers carrying the Asrock model so I sent them an ask when its expected here and they responded with "we will not be selling it in America". I do worry about getting it in from Ali express but also it is way more expensive than it should be because of rarity.
smallformfactor.net
Last edited:
How is your RAM configuration on this system?
I wouldn't expect too much from the 8500T but with recent CPUs Intel improved their iGPU performance a lot.
Edit: my poor results for 3200G is related to the A320TM-ITX board which only runs on single channel RAM despite AsRock saying it would support dual channel:
Motherboard - AM4 Thin ITX motherboards, They do exist!
I can't find any US retailers carrying the Asrock model so I sent them an ask when its expected here and they responded with "we will not be selling it in America". I do worry about getting it in from Ali express but also it is way more expensive than it should be because of rarity.
I have 2x8GB Flare-X at XMP3200, single channel IIRC. The 8500T ran about 5.8FPS although I didn't go to the end I don't think it would've changed that much, but I also tried it using openGL to about the same point and that was <1 FPS slower than direct3D11(8500T memory 2x8GB @2667MHz). My 2400G is capped at 65W and the 8500T is 35W. The Intel iGPU actually did fairly well considering the CPU/GPU/memory clock speed disadvantage.
I think you are running dual channel, see the 8x in your results and the 4x in my results of the 3200G. Not sure why the 4650G is showing 6x with the JUBO A300 though.
I'll definitely have to re-test my APUs using another board. They seem to run quite crappy on thin mini ITX.
You're right, just went and looked up, dual for sure. shouldn't have been so lazy. Maybe thinking single rank.
Last edited:
