After using a high end 3900x build it’s safe to say AMD is doing pretty well. I expect for the 3950X to be on par or above the 9900k for users that manage to get their IF clocks to 1:1 with 4K MHz RAM. It’s difficult but it has been done with the right overclocking. CL14 3800 MHz has already shown large gains for anyone using RAM intensive applications.
My FPS in Total Warhammer 2 doubled over my old 4970k build and I use a 4K/144 Hz screen. I have actually hit 60 FPS with everything maxed except AA.
Neat, i wonder if the 9900KS will clock to 5.1/5.2/5.3 somewhat easily like the 9900k can do 5ghz. Either way it looks like i'll be waiting for 2020 to upgrade from my 7700k/gtx1080. That's when some worthwhile hardware will be launching with zen 3 and whatever intel musters up. Going to be interesting what AMD has to dethrone the 2080Ti and Nvidia's counter.
Hardware Unboxed is just the YouTube version of the same content in Techspot. It's actually sad to see people complaining about how content is delivered but fail to recognize the same YouTubers also have articles on their own websites. Shouldn't people be happy that there is more than one option to consume the same content?
Potayto potahto. Stating that one is better than the other for anyone other than yourself requires an assumption that everyone else is in agreement with your criteria for what makes a youtuber good/bad.
For some people, personality is way more important than knowing how to use an oscilloscope. For other people, extensive knowledge of an oscilloscope is the key to their heart. Those two people are not going to be in agreement on which tech youtube channel to watch, but it doesn't make either one of them wrong or right.
The complaint is quality of youtube videos - test methology, not well documented if that at all, not knowing how to test things or use equipment to test.
Go watch a Igor'sLAB video or check out his website then show me a Techspot/Hardware Unboxed video with that level of quality of content.
Besides have lab equipment and actually knowing how-to use it. Having a multimeter and knowing how to use it.
Now all the big youtuber are buying thermal image camera which Igor Wallossek has been doing for year. Watch them make a lot of mistakes like when GPU reviewers are trying to do motherboard reviewers.
I won't refute that, Igor has done some great work for Tom's Hardware and his own site although I'm not a German-speaker, so I automatically don't consume his content.
I'm not sure why people are expecting the same quality from YouTubers as the detailed reviewers like Igor, VSG or jonnyguru. It's quick mainstream content vs lab-tested environments, but they are all essentially free content. We aren't directly paying them to do a review for us, so we aren't entitled to complain about how/what they test and publish.
Regardless, I don't see the issue with these more popular YouTubers. They do report their test setups and updates to those setups but it's only listed once in a while as my assumption is that their followers will be watching their videos and will get annoyed to see the same test methodology shown in every single video.
People on enthusiast forums know better then the general public. And everybody is entitle to enjoy things. But I have a major issue when these
Youtuber's show 1.40 volts for Ryzen 3000 being a safe overclocking voltage implicit or not. Their audience see's graphs with 1.4v for Ryzen 3000 they
rightfully assume that it is safe and that they can run that 24/7. It is beyond what previews generation Ryzen (Pinnacle Ridge) 12nm could do before degradation occurred.
If you are going to call yourself a Tech Journalist then you should conduct yourself as such. 1.40v is more than I had my Westmere 32nm CPU on.
Hardware Unboxed destroyed their first 3900x leaving voltage on AUTO and messing around with AUTO overclocking. Didn't stop them from doing AM4 motherboard reviews with 1.40v.
GamersNexus also has video and on their website showing 1.40v for Ryzen 3000 overclocking.
Intel overclocker might not be familiar with The Stil's work and how much he has contributed to the AMD community. Below is what he had to say about Maximum Safe voltage for 7nm Ryzen (Matisse)
"According to FIT, the safe voltage levels for the silicon are around 1.325V in high-current loads and up to 1.47V in low-current loads (i.e ST), depending on the silicon characteristics" - The Stilt
Essentially, if we're talking about the higher-end SKUs, there is basically none.
Based on my experience, the best case of scenario on 6C CCDs (3600, 3600X and 3900X) is around 4.25GHz, at relatively safe voltage levels.
In case of 3900X, given that you can cool the chip with two of those 6C CCDs. SKUs with 8C CCDs (3700X, 3800X and 3950X) the best case is around 4.15GHz. The 3950X is expected to be thermally limited, as a whole.
The biggest limit is the intensity (heat per area), secondly the voltage you can safely feed to the silicon. For example, the 9900K which has a reputation of being an inferno, has theoretical intensity of ~1.15W/mm² when operating at 5.0GHz (200W @ 174mm²).
Meanwhile Matisse can easily reach intensity of > 1.5W/mm² (120W+ @ 74mm²). The second issue is, that beyond ~3.8GHz the V/F curve becomes extremely steep. According to FIT, the safe voltage levels for the silicon are around 1.325V in high-current loads
and up to 1.47V in low-current loads (i.e ST), depending on the silicon characteristics. Because the stock boost operation is already limited by the silicon voltage reliability, the only way to eke out every last bit of all-core performance is using OC-Mode. Like on previous Ryzen generations, entering OC-Mode also means that you will loose the turbo boost (all cores operate at same frequency). On the higher-end SKUs, the single threaded performance penalty will be massive from doing so. For example on 3900X, you'd be trading additional ~100MHz all-core frequency to a loss of up to 450MHz in ST frequency by doing so. Personally, I advice against overclocking the higher-end SKUs at all, and instead increasing the power limits and trying your luck with the "Auto OC" feature (which most likely isn't beneficial).
The V/F testing was done using full resource utilization (FRU), meaning the stability was tested using 256-bit workloads.
Unlike Intel designs, Matisse does not feature an offset for 256-bit workloads. This means that to ensure the stability of the CPU cores in every scenario, they must be tested using this kind of a workload.
On Matisse, the delta in power consumption between the scalar and 256-bit vector instructions is massive, as expected (37%). That being said, there seems to be other design related factors limiting the maximum achievable frequency.
Despite significantly lower power consumption and therefore also lower temperatures, stability even in pure scalar workloads could not be achieved at much higher frequencies, compare to FRU scenario.
Performance per Watt
As expected, Matisse provides significantly higher performance per watt than its competition, thanks to its leading edge 7nm manufacturing process. Some of you might notice that Matisse's power efficiency seems to peak at 3.5GHz, despite the fact that semiconductors do not behave like that. The reason behind this was revealed by Vmin testing, which clearly illustrated that Matisse lacks fused V/F (voltage-frequency) curve below 3.4GHz. This means that below 3.4GHz frequencies the voltage is always at the same level, it is at 3.4GHz. The stock (fused) V/F curve appears to be extremely well optimized as well, leaving only the temperature factor on the table.