FM2+ Athlon X4 860K in Dolphin benchmark

[rafaelluik]

Guys I'm really curious about how this CPU performs in Dolphin, as there's some IPC (single-thread) and overall (multi-thread) improvements over its predecessors.

Could anyone test it in the Dolphin benchmark?
https://forums.dolphin-emu.org/Thread-new-dolphin-cpu-benchmark-no-game-required

Please use the latest "developer version" (it doesn't require installation) https://dolphin-emu.org/download/

I'd be interested to know also how playable are a few games like Super Mario Galaxy, Zelda Skyward Sword and Metroid Other M just to have an overall picture in my head. FPS... Although anything below 60 FPS will just break the sound since it's synchronous right?

[KHg8m3r]

You would be better off getting a G3258 and overclocking to around 4.7 GHz.

As for how well it performs, it will probably perform (at stock) slightly better than the Athlon X4 750K, which still only puts it at the lower end of the Dolphin benchmark: https://docs.google.com/spreadsheet/ccc?key=0AunYlOAfGABxdFQ0UzJyTFAxbzZhYWtGcGwySlRFa1E#gid=0

[rafaelluik]

I know about the Pentium, but in the Intel side I wouldn't go lower than an i3 41XX really so the machine is not tied to a CPU that can only handle 2 threads and that's going to produce more heat, require a fan or something for OC, decreasing the part lifespan, etc...

750K is not representative here, it's an old Trinity processor while 860K is a Kaveri (2+ generations newer) that would be somewhere near an A10-7850K (with a dGPU that won't bottleneck either of course).

[KHg8m3r]

Problem is that AMD hasn't invested into increasing their single-core speeds beyond going down to a smaller fab size, so there won't be some sore of magical speed boost in Dolphin.

Well in that case it would probably be around the A8-6600K or FX-8350 from what you're saying. That's still not great though.

[DatKid20]

860k would actually have the exact same IPC as the a10-7850k. It would however be faster then the one in the benchmark due to higher clocks. If you're going to buy a system for this remember to buy one with quality vrms. Most peoples VRMs get super hot around 4.5ghz.

I checked the benchmark and a 4.7ghz Kaveri chip could handle most games at full speed.

[NaturalViolence]

rafaelluik Wrote:I know about the Pentium, but in the Intel side I wouldn't go lower than an i3 41XX really so the machine is not tied to a CPU that can only handle 2 threads and that's going to produce more heat, require a fan or something for OC, decreasing the part lifespan, etc...

Are you implying that not having HT will somehow increase heat? This makes no sense.

The G3258 is a better OCer than any i3. It is cheaper and if OCed much faster. You have no reason to insist on getting an i3 or higher.

rafaelluik Wrote:Guys I'm really curious about how this CPU performs in Dolphin, as there's some IPC (single-thread) and overall (multi-thread) improvements over its predecessors.

This sentence also makes no sense. IPC effects overall performance (both single-threaded and multi-threaded), not just single-threaded performance. And "overall" performance is tied to both single-threaded and multi-threaded performance, not just multi-threaded.

[ThorhiantheUltimate]

To further clarify what Natural Violence said, Clock speed multiplied by IPC (which stands for instructions per cycle) = IPS (Instructions per second). To make things simple, each core is going to have an output of some measurement of IPS. That measurement would be the "single-core" performance, and the performance of each individual core will probably be similar to each other. The actual total performance that can be done by using multiple or all of the cores on a CPU would be a measurement of "multi-core" performance. Since this is basically a sum of the IPS, or whatever unit of measurement you use in a benchmark, throughout the multiple cores in the CPU, raising the clock speed or, as stated by NV, the IPC of each core will raise multi-core performance since each individual core will have a higher IPS.

[rafaelluik]

(11-06-2014, 11:58 AM)NaturalViolence Wrote:

rafaelluik Wrote:I know about the Pentium, but in the Intel side I wouldn't go lower than an i3 41XX really so the machine is not tied to a CPU that can only handle 2 threads and that's going to produce more heat, require a fan or something for OC, decreasing the part lifespan, etc...

Are you implying that not having HT will somehow increase heat? This makes no sense.

Nope, I'm saying that OCing Pentium = more heat, worrying about getting the OC right, decreasing its lifetime, an aftermarket cooler... Just get the i3 41XX and you're done! And, on another note, when you're using the PC for things other than Dolphin the extra threads will be more worth it.

(11-06-2014, 11:58 AM)NaturalViolence Wrote:

rafaelluik Wrote:Guys I'm really curious about how this CPU performs in Dolphin, as there's some IPC (single-thread) and overall (multi-thread) improvements over its predecessors.

This sentence also makes no sense. IPC effects overall performance (both single-threaded and multi-threaded), not just single-threaded performance. And "overall" performance is tied to both single-threaded and multi-threaded performance, not just multi-threaded.

I understand that, I was just saying this particular processor sees improvement in the two areas compared to its predecessors. You can include IPC under "overall" (together with the other factors that improve multi-thread performance) when you read.

I feel that I wrote it that way because better IPC not always lead to better multi-thread performance because there are other factors. If we were to compare the i3 4150 to an Athlon X4 860K, according to PassMark they switch places in the two scores.
Average CPU Mark - i3 4993, Athlon 5761
Single Thread Rating - i3 2045, Athlon 1581

[Link_to_the_past]

(11-06-2014, 10:15 PM)rafaelluik Wrote: I feel that I wrote it that way because better IPC not always lead to better multi-thread performance because there are other factors. If we were to compare the i3 4150 to an Athlon X4 860K, according to PassMark they switch places in the two scores.
Average CPU Mark - i3 4993, Athlon 5761
Single Thread Rating - i3 2045, Athlon 1581

It is due to the core number.