https://www.youtube.com/watch?time_continue=2&v=Bc3qjnnX1PA
>>59748625
You ain't getting any views from us.
chink shill.
>inb4 intel shill
>Internet explorer on Windows 10
An youtube video will kill my mobile data. What is it about?
>>59748707
Watch it, it's worth it
>>59748707
A dude with the new windows 10 creators update 1703 with overclocked ryzen 1700 to 3.8ghz has interesting benchmark results.
>>59748731
is next week 1703? I thought it was 16something something 8?
just post the benchmark compared to the previous version
>>59748805
This
>>59748805
This
>>59748707
see screenshot in op, thats basically it
>>59748625
I thought it was the infamous sleep bug first, but aida64 timings aren't off. what the fuck is going on?
>>59749137
maybe it's ... dare i say... a good CPU?
>>59749162
it was already to begin with. but 20% improvement with just a patch seems strange.
I really hope others can replicate this.
>>59749217
There's something finky going on with Windows and Ryzen, since Linux can exhibit a lot higher performance in certain tests, it's either some ACPI, P-state or scheduler update on the Windows side.
Interesting to say the least, will love to see real applications benchmarked instead of just benchmarks, will be nice to see if it's an overall update or if its a they optimized for benches update.
>>59748625
Fuck off AMD shill
>>59750022
Fuck off pajeet, you aren't getting paid for your posts
>>59749959
synthetic benchmarks tell you more about the full capabilities. of course this ultimately has to be translated into program performance to be of real-world relevance, but the specific programs used to benchmark are more susceptible to selection bias and specific idiosyncrasies which paint an otherwise good performing CPU in a bad light.
>>59750049
Oh no shit, but the point is, if this is a per application optimization and not an across the board one, then its shitty as it just makes it look like amd is cooking the benchmarks.
No way this real, at 4.0ghz cinebench score 1800.. how possible to get 2000 at 3.8
>>59750083
Look for yourself. This cinebench multi-rendering takes 20 seconds at 3.8ghz. Watch any video of overclocked ryzen cinebench and you'll see that everyone renders in about 24-26+ seconds. Here is an example: 4ghz and 2933Mhz ram, 24 seconds https://www.youtube.com/watch?v=8IuzF73zMJg
>>59750069
This result is after a windows 10 creators update. It's not a per application optimization. And, even if it were a per application optimization, it would still mean ryzen cpus are capable of more.