Also, your link is biased because it is well documented that AMD performance significantly improves with faster RAM and they used much faster RAM in the test only for AMD.
That RAM is very expensive so you need to factor that in the total cost if you wanted fo be fair when doing price comparison as well.
Strongly disagree. Zen 2 gave AMD better IPC and comparable clock speeds. Zen 2 is also more power efficient than what Intel offers per unit of compute.
Intel offers a few processors with higher frequencies, but only by having absurdly high power consumption and uncompetitive price.
Source: https://www.anandtech.com/show/14605/the-and-ryzen-3700x-390...
Look at the last chart at the bottom and the summary:
> Normalising the scores for frequency, we see that AMD has achieved something that the company hasn’t been able to claim in over 15 years: It has beat Intel in terms of overall IPC.
The web tests are also very lightly threaded benchmarks:
https://www.anandtech.com/show/14605/the-and-ryzen-3700x-390...
Definitely no “ownage” there from Intel. AMD processors did great. The Intel 9900K generally did worse, and the 9700K only did slightly better for some interesting reasons discussed at the end of the page.
Making claims that AMD’s processors only perform better in highly threaded tasks is either disingenuous or based on significantly outdated information.
Zen 2 brings AMD up to parity with Intel in IPC, and the clock speed difference isn’t substantial in most cases.
AMD also offers substantially more (equally strong) cores and stronger threads at a better price. They also have products that offer substantially more PCIe 4.0 lanes than anything Intel has, among other nice features, especially when comparing server processors.
Intel will dump billions and billions of dollars into becoming competitive again, but right now... Intel’s main advantage is in their low idle power consumption in laptops. No one should be buying Intel processors for anything else.
This was more true in prior generations of AMD. Newer ones aren't as RAM dependent (largely due to increased L3 cache and a more efficient core interconnect topology). Of course any CPU will still perform some amount better with better RAM.
> and they used much faster RAM in the test only for AMD.
This is a bit of Intel's own fault. AMD officially supports 3200 MHz while Intel officially only supports 2666 MHz, anything higher is considered overclocking the memory controllers even though both are perfectly capable of going much higher than that. Also note the Intel has quad channel while the AMD has dual channel. Quad channel has higher bandwidth but is the same RAM is going to have to run at lower speeds that if it were dual channel.
> That RAM is very expensive so you need to factor that in the total cost if you wanted fo be fair when doing price comparison as well.
The link to the bench now shows "not found" so I don't know what the exact RAM used was but you can get 16 GB of well timed 3600 MHz RAM for $80 so whatever difference there is isn't coming close to affecting that $250 price gap.
Intel certainly does still have an edge on single thread performance but the interesting thing is it's an ever shrinking lead at an increasing cost
Go buy whatever you can even if its subpar these days, nobody else cares, but currently most performance is with AMD. Power per dollar in high performance is purely with AMD.
It depends on the pricing tier.
My 3700X costs about 20% less than a 9700K but is a bit faster in single core, and about 30% faster in multi core.
https://www.cpubenchmark.net/cpu.php?cpu=AMD+Ryzen+7+3700X&i...
https://www.cpubenchmark.net/cpu.php?cpu=Intel+Core+i7-9700K...
Real-world test with mitigations:
https://www.phoronix.com/scan.php?page=article&item=3900x-99...
Or https://www.techpowerup.com/review/amd-ryzen-7-3700x/22.html
9700K is still solidly faster in per-thread performance. Having SMT gives the 3700X a boost in total performance though. The 9900K beats all the 8C Ryzens (including 3800X) and beats the 3900X in per-thread performance but loses in total performance (of course).
In some games the 9700K seems to have better performance, but in pure CPU benchmarks the 3700X is generally faster.
See these results: https://www.techquila.co.in/amd-ryzen-7-3700x-vs-intel-core-...
In Cinebench the Ryzen has the advantage in both single and multi core.
See this other results for real time audio performance where the Ryzen is 10-15% better in all results: https://www.scanproaudio.info/2019/07/12/amd-ryzen-3600-3700...
Also, in your second link the price-performance-ratio of the 3700X is 27% better...
If your workloads are limited by single thread performance you need better software. It's why vulkan and dx12 are a thing. (The single thread limitation of committing a frame to the GPU has been reduced by an order of magnitude) It's why C++ has the parallel algorithms library baked into the language.
I get it, threading is hard. But it's honestly not that hard. It's only hard when you're maintaining some super old program with single threadedness engineered into its core architecture. (note: this is my day job) Greenfield applications since 2009 should have had threading built in as a core assumption.
AMD is doing the right thing by optimizing for multithread performance over single thread performance. Moore's Law is dead for single cores. It has been for a decade and a half.
However, Amdahl's law means that serial/single-threaded performance remains an important determinant of total performance even in multi-threaded applications. Locking is a great example, even in a heavily multithreaded program you can be limited by lock contention, that specific part of the program is serial.
Gustafson's Law is of course a thing, eventually we will come up with additional work to fill those cores, but single-threaded performance remains a critical factor for CPU performance.
Most of the time we don't get to choose which software we run. Hell, most core components of Windows get stuck in a single core still. I'll take single-core performance over multi-core performance any day. Maybe that will change in 10 years, but I live in the present.
If looking at the geometric mean for these various mitigation-sensitive benchmarks, the default mitigations on the Core i9 9900K amounted to a 28% hit while the Ryzen 7 2700X saw a 5% hit with its default Spectre mitigations and the new Ryzen 7 3700X came in at 6% and the Ryzen 9 3900X at just over 5%.
https://www.phoronix.com/scan.php?page=article&item=amd-zen2...
Should note:
Keep in mind these benchmarks ran for this article were a good portion of synthetic tests and focused on workloads affected by Spectre/Meltdown/L1TF/Zombieload. Many of these particular tests aren't multi-threaded and that's why you don't see as much of a difference between these HEDT and desktop CPUs as in our more normal benchmarks.
That isn't always true, for instance on some CS:Go benchmarks, the new Ryzens were quicker, and that is dominated by single threaded CPU perf.
Also, gcc compile time benchmarks are faster on the new Ryzens, largely due to large L3 cache sizes on the Ryzens (which is facilitated by their chiplet design)
It would be more accurate to say that intel has a slight single thread performance, most of the time. Rather than to say "owned".
Its all close enough that it depends on lots of details, like cache size and ram speed and timings. (ryzen benefits more from ram speed due to the interaction with infinity fabric). You also have to consider whether mitigations for side channel attacks are on for each chips.
There are dozens of little details like this that can be taken into account when deciding on a benchmark, and there is not always a clear answer as to "what is fair". You just have to be sure you say how those parameters are set, so people know what the comparison is, exactly. You can't be so quick to complain of bias, just because the parameters are not set the way you would have wanted them set.
I thought the RAM cartel backed off the ridiculous prices when the Justice Department started murmuring about antitrust.
What really got them was when China stepped up investments in a number of projects to jumpstart some domestic fabs focused on RAM production. The RAM cartel came to the table and signed some agreements regarding RAM production quotas with the Chinese real quick after that.
https://semiengineering.com/will-china-succeed-in-memory/
https://seekingalpha.com/news/3327772-samsung-headed-china-a...
A huge percentage of China's economy depends on smartphones, laptop, desktop, server manufacturing and they had no interest whatsoever in seeing that drastically slowed down so that Samsung, Hynix, and Micron could enjoy a fatter margin.
Not like a majority or anything, but losing 10-20% off 10% of your economy is a big deal.