3 ms·
The improvement in the Dolphin benchmark (which probably indicates improvements in other JIT-based emulators/runtimes) is pretty impressive, especially compared
by 0xcde4c3db 7y ago
The improvement in the Dolphin benchmark (which probably indicates improvements in other JIT-based emulators/runtimes) is pretty impressive, especially compared to first-generation Ryzen. For a while that benchmark had every AMD processor performing worse than every Intel processor in a given review.
If the Chromium compilation benchmark (which reportedly was screwed up and will be added later) shows similar improvements, I think this generation will finally convince me to replace my Ivy Bridge box.
edit: I just saw the Phoronix benchmarks, which show LLVM compilation going from 547 seconds on 2700X to 406 seconds on 3700X. I'm not sure whether those were run with the same chipset/RAM/storage versus the 2700X result being an older one pulled out of the database, though.
- phire 7y agoDolphin is probably very happy with the improvements to branch prediction. As for compiling, Gamers Nexus' review of the six core 3600 had it absolutely trashing other CPUs (33% faster than the 9900k, 11% faster than the 2700x) in their GCC self compilation test.