Hello youtube, got a quick one for you today. We're doing caveman overclocking on my R5 1600 in the hopes of getting 4GHz out of it.
Quick history lesson if you didn't know this was from the Summit Ridge Zen midrange from 2017 so the first AM4 cpus and using a 14nm node. The 1600 in particular was the third lossiest yeald of this generation overall boosting only up to 3.7GHz and only beating the 12 and 14 hundread in this aspect whist getting a worst extended frequency range than the likes of the 1300X and on top of all of that these chips don't OC for shit above 3.8 and there is very little to be gaind above that anyways. So as we're about to get into it lord have mercy on this CPU.
Today we're running Cinebench R23 because 26 R24 just refused to launch; on top of the in house 900MB Windows 11 26200.7705 as in Windows 11 that idles at 900MB dram with bearbones scheduling. Stock the CPU gets a multicorescore of 6200 and a singlecore score of 940. I also forgot to mention that it's mounted on a MSI B350m Bazooka motherboard. You can check out the performance difference from stock to 3925MHz in the are six cores enough in 2025 video as we will continue from there. Also in that video, the cooling solution is explained basically, we're running the CPU delidded with 128W/mK liquid metal TIM under a Thermaltake 90mm radiator with a 2 120mm fan push pull configuration.
So, at the 3925MHz checkpoint, we are running plus 0.1875 volt as an offset and 1.1125 volt SoC voltage and that gets us 7588 multicore score and 1019 singlecore score. This also results in a 107W potntial power draw and 56.5 degrees on average with 57.5 dergees max on the CPU as we moved to an ambiental tempreture between 0 and 5 degrees celcius. Here are both the runs, sorry for the chopped image, I just didn't bother with graphics drivers for this one.
From here, we moved on to 3950MHz with the same voltages and that returned 7630 multicore score and we saw on average 54.5 degrees and 55.5 degrees maximum on the CPU probably because the ambiental dropped a bit and also a 119W spike for package power draw. Here is the run:
After a few attemts we also got 4GHz working with plus 0.25volt offset and the max SoC voltage the moitherboard allowed of 1.3volt. This was probably more than it needed refering to SoC but after it didnt work with the 1.1125 i just maxed it out and called it a day. The funny thing with 1.3 volt SoC is that sometimes the USBs wouldnt work and since i dont have PS/2 peripherals i had to restart each time until they decided to work. Anyways 4GHz got us 7726 multicore score and 1042 singlecore score with temperatures of 55 degrees celcius average and 55.5 degrees max and a power draw spike of 121 watts that is close to double the CPUs stock TDP so safe to say this CPU was never designed to run anywhere near these specs. These are the runs.
Unfortunatelly 4025MHz would not be stable under any voltage configuration so I can safely say that we achieved this particular's CPU limit at 4GHz. So where does that leave us and what can we compare it to?:
The short of it is that we increased multicore performance by 24.6% and singlecore performance by 10.85% from stock and this comfortably beats a stock 10400 or 5600U or 2600X in mulricore score and is getting very close to a stock core 5 120U. In terms of singlecore score we're beating a stock 6600K or 2400G while nearing the performance of a stock 10310U.
And that's basically all I wanted to yapp about in this video, sorry there isn't more to it but this chip just wouldnt budge above 4GHz and already required housefire levels of voltage just to go above 3.8 so, yeah. Anyways hope you enjoyed and thanks for watching.