AMD Zen 3 Ryzen Deep Dive Review: 5950X, 5900X, 5800X and 5600X Tested
by Dr. Ian Cutress on November 5, 2020 9:01 AM ESTCPU Tests: Synthetic
Most of the people in our industry have a love/hate relationship when it comes to synthetic tests. On the one hand, they’re often good for quick summaries of performance and are easy to use, but most of the time the tests aren’t related to any real software. Synthetic tests are often very good at burrowing down to a specific set of instructions and maximizing the performance out of those. Due to requests from a number of our readers, we have the following synthetic tests.
Linux OpenSSL Speed: SHA256
One of our readers reached out in early 2020 and stated that he was interested in looking at OpenSSL hashing rates in Linux. Luckily OpenSSL in Linux has a function called ‘speed’ that allows the user to determine how fast the system is for any given hashing algorithm, as well as signing and verifying messages.
OpenSSL offers a lot of algorithms to choose from, and based on a quick Twitter poll, we narrowed it down to the following:
- rsa2048 sign and rsa2048 verify
- sha256 at 8K block size
- md5 at 8K block size
For each of these tests, we run them in single thread and multithreaded mode. All the graphs are in our benchmark database, Bench, and we use the sha256 and md5 results in published reviews.
GeekBench 5: Link
As a common tool for cross-platform testing between mobile, PC, and Mac, GeekBench is an ultimate exercise in synthetic testing across a range of algorithms looking for peak throughput. Tests include encryption, compression, fast Fourier transform, memory operations, n-body physics, matrix operations, histogram manipulation, and HTML parsing.
I’m including this test due to popular demand, although the results do come across as overly synthetic, and a lot of users often put a lot of weight behind the test due to the fact that it is compiled across different platforms (although with different compilers).
We have both GB5 and GB4 results in our benchmark database. GB5 was introduced to our test suite after already having tested ~25 CPUs, and so the results are a little sporadic by comparison. These spots will be filled in when we retest any of the CPUs.
Facebook
Twitter
RSS
339 Comments
View All Comments
Badelhas - Friday, November 6, 2020 - link
I totally agree. I've upgraded from the last true overclocking champion from Intel (i5 2500k @4.8ghz from 8 years ago) to the 3600, it was finally worth it but going from 200 to 300 euros is a bit to much of an increase in price, in my humble opinionSpunjji - Sunday, November 8, 2020 - link
They're not really comparable, though. I'm weirded out by how many people are comparing the 3600 to the 5600X. The X is a bit of a giveaway.Kallan007 - Saturday, November 7, 2020 - link
I just buy new and sell off the old. But if you want a price break then just wait.Spunjji - Sunday, November 8, 2020 - link
I doubt it will. They'll sell every one they can make, and if not, there's no reason they can't begin to lower prices as supply begins to exceed demand.Threska - Monday, November 16, 2020 - link
Socket longevity is the important thing here for anyone playing the value game. You may not buy the latest and greatest NOW, but the future allows for it without starting completely over.UNCjigga - Thursday, November 5, 2020 - link
I suppose the only thing missing is a chipset/IO package with USB 4 support? Not a big deal for desktops--but I hope they have that figured out by the time Zen 3 is ready for mobile parts.Spunjji - Sunday, November 8, 2020 - link
That would be nice to see. I have a suspicion we won't see it until the new socket arrives on desktop, but would be good to get it with Cezanne on mobile.Machinus - Thursday, November 5, 2020 - link
Looks like a great set of chips for anyone who gets one mailed to them directly from AMD.Good luck buying one in a store.
charlesg - Thursday, November 5, 2020 - link
I have to say I'm disappointed in the availability of the 5900 and 5950. I expected better.lmcd - Thursday, November 5, 2020 - link
Yea honestly isn't this the whole point of the chiplet model? Or is the IO die different for the 2-chiplet models? I assume it's not packaging constraints because that makes no sense.