Oh, the 9800x3d does perform “better”, but** STILL** a bottleneck 
ead:
But wait, the 5950X is better!
Here’s the equalizer - a 14700K beats the 9800X3D :cigar:
1 Like
Ok bought the 5080 also. So hope 5700x3d + 5080 combo will be a good one till maybe AM6.
3 Likes
If only AMD launched 5950x3D, id have bought it without hesitation. OP should get 10-20% improvement overall and more on the 1% low with 5700x3d for sure.
Site owner must be one of the userbenchmark fan.
Congrats. Don’t forget to undervolt both processor and the GPU.
Thanks. Will definitely do it as I saw the improvements in my existing setup
1 Like
I have a question. I have a Samsung galaxy book pro 4 360 with an oled panel, if I stream from my main pc using Apollo/Sunshine and moonlight, will the picture quality be same as if I was running an oled monitor directly to my main pc. This will be over ethernet with a bit rate of around 100 Mbps. It will help my decision to get an oled if I can see the performance beforehand. Thanks
Try it for yourself, usually latency & picture quality are issues with streaming.
It has HDMI 2.1 so you can directly attach to it from the PC case.
My understanding was that the hdmi port only does display output to other display.
I have tried it, my query is whether the streaming at 100 mbps , the picture quality is representative of how it would look on a OLED display if it is connected directly.
Nah you can use it as main output as well. In fact newer HDMI is better than Displayport.
It should. Also you will be using the WIFI if you are doing local streaming. So your internet speed won’t matter.
I meant the HDMI 2.1 port on my laptop is output only. It wont display my Main pc.
My laptop is also connected through ethernet.
That sucks, I didn’t get it at first. Streaming through ethernet will be rock solid (that is assuming Apollo/Sunshine/moonshine etc supports it which i think they do) but like @@OMEGA44-XT has said, you will have to try it.
Upgrade to 5700x3d + RTX 5080 done. 2x framerate in Star Wars Outlaws . 3x with 2x Frame Gen.
My CPU timespy score is 10173 which is lower than what I saw online. If anyone that has this can confirm their score?
Also undervolted 5700x3d -30 all cores.
RTX 5080 undervolted and overclocked to 3000 Mhz at 20 watts below stock power draw.
1 Like
I got 11205 CPU score on first run where GPU was limited to 70% power and 10883 on second run where GPU was at 100% power. Not sure why lower score on second run because average clock frequency stayed the same for both runs (4,039 MHz) although temp was a few degree higher on second run. -30 all core UV.
Your CPU seems to be slightly under performing indeed. Can you check and tell me what was the average clock frequency and temp?
Screenshots attached.
Frequency seems to be locked at 4049 Mhz and the temps were mid 50s with the end spike to around 59C
Why ? Use the tool like “Hydra” to calibrate and set CO per core basis to achieve the best single/multi core perf OR just leave them on stock PBO.
will try hydra and see.
Also without the undervolt, I was getting similar numbers so thought will try -30 all core and was stable in corecycler so left it. Atleast get lower power usage and temps with the same performance.
Did you try any CPU intensive games yet?
Why not? Why use “hydra” when you got native undervolt support in bios. Also 5700X3D is very easily able to do -30 all core UV without stretching and/or crashing. AMD really crippled the Ryzen 5000 X3D chips with locking their max frequency. The amount of headroom and OC potential they have is insane. Using per core undervolt to achieve best single/multi core performance would have been understood if the chip had capability to push further than the locked 4.1ghz. If it’s able to achieve the same frequency (max) on -30 all core without stretching (and/or crashing) then it shouldn’t affect it’s single/multi core performance. From the looks of it his processor’s -30 is working fine if we look at the average clock frequency at 4049mhz (4.1ghz). In fact the single core boost is working fine as well by looking at the bump in the graph which is/should be at 4150mhz.
Of course OP can try hydra or just set undervolt to -20 and see if this improves the benchmark performance since there is a big difference right now.
I tried black ops 6 benchmark tool and it shows a 10 % bottleneck. Other than that Star wars outlaws has cities which seems to be very heavy on the processor and in that I have good enough performance. Any other games you guys can tell and I will check.
I checked hydra and reddit threads are full of people saying it fried their processor and half the people saying it worked for them. I didnt find it worth the risk to do.