Principled Technologies Issues Updated Test Results

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Actually... TOM'S HARDWARE TESTERS... How about a performance-for-equal-priced-systems test?

You know, pit the optimal video card, optimal memory, optimal motherboard, optimal SSD against each other ... do NOT make them "the same", just make the overall price of the system(s) the same.

Let Intels's "optimal performance" system lead the CASH end...
Then tune AMD's system accordingly. Within 1% of the same price.

I'm pretty sure then the advantage won't be Intel's.
Just saying.
 
Considering that there's 0 mention of modifying the RAM XMP profiles (which in combination with Creator mode give better performance on Ryzen) and no apparent change in CPU cooler (which could be preventing the 2700X from reaching higher/proper boosts), these 'updated comparisons' don't really mean much.

At any rate, I don't think getting the i9-9900K is worth the price vs the 2700x.

Furthermore, with soon to be released Zen 2, we're looking at 8c/16th Ryzen 3700 (for example) with base clocks of about 4.35GhZ and boosts beyond 5GhZ on single or two cores....
With all core boost being around 5GhZ likely at 65W.

The 95W version (3700x for example) will likely have a base clock speed of 4.93GhZ (aka 5GhZ), and unknown boosts across single and all cores (though I wouldn't be surprised if AMD can push it to 6GhZ considering the 7nm TSMC node is better than Intel's 10nm and is designed for high performance and efficiency).

Plus, 45% performance increase was touted... that's performance, not clocks... and since Zen 1 needed about 15% clock boost to achieve 10% performance boost, Zen 2 would need to be clocked 60% higher to achieve that... which is not impossible... but I'm prone to give Zen 2 a more conservative 45% clock increase over Ryzen 1 with another 10-15% (potentially) coming from IPC/architecture improvements... at same TDP's of course.

So, at the very least, the 65W version of Zen 3700 for example should match Intel's i9-9900K in clocks, if not slightly surpass it in clock speeds... whereas the 3700X (95W TDP version) is likely to go well beyond that.
 


The RAM speeds were both set to the maximum officially supported speed for each platform. 3000MHz memory down clocked to 2666 for Intel, 2933 for AMD to avoid having anything overclocked. This was the setup used for the original tests as well though not properly disclosed in the benchmark notes.
 


Yes, there are specific hardware optimizations in Premiere that are exclusive to Intel IGP's and can speed up rendering by about 30% depending on work load on top of CUDA acceleration. They can be used at the same time. Gamers Nexus and Hardware Canucks both did Youtube articles on this. You don't have to disable your dGPU to use IGP video encoding acceleration. Not sure what you are talking about there.

Just to add, CUDA can accelerate only a select few operations:

some effects (complete list at the bottom of this post)
scaling (details here)
deinterlacing
blending modes
color space conversions

You would not buy a highend GPU for Premiere to speed up rendering.
 


The proble is that the 9th gen is having new motherboards. Even if you have most of your system, you are trap with an expensive motherboard and the most expensive 8 core CPU at value above the MSRP by 50-100$.

Basically, you can almost get a 2700x + a 1080 TI and a b450 motherboard for the same price. If you add the cooler, it is even worst.
 


Okay, let's not go there. The HEDT platforms are not RELEVANT with gaming benchmarks. They exist for computing and production. It is nice to see the result, however when you compare the more expensive HEDT from Intel, they don't fare that much better.

It is still a 300$ CPU compared against a 550$ CPU and a 80$ HSF. For getting a mere 10-15% at 1080p with a 1080 TI at medium settings, you know that Use Case is utter BS. Nobody in their right mind with that hardware will do that.

They will play at 1080p ultra @ 144Hz, 1440p ultra @ 75Hz or 2160p @ 60Hz. Basically, in all these scenarios, the 9900k is having nothing justifying the 550$ (630$ with HSF) price tag on the street... absolutely nothing.
 
I think they only published retest results for the AMD with stock cooler at the insistence of Intel, I bet the margin would have been smaller yet, and the smaller margins, even though Intel would still be ahead, scares Intel.
 
Intel is probably pooping their pants right now. Being stuck on 14nm longer than they should, having shortages (correct me if im wrong). Well I guess my next path is a 2700x or a 2800x and pair it with a 2080Ti for some 4k gaming.

still, personally i find it funny how they compare 2 CPUs one with box cooler and the other with literally one of the best air cooling solutions in the market right now and the 2700x is only ~10% slower. and call it fair. hahahaha

Intel is so scared they have to rely on these weird tactics to throw people off and "show" reasons why their processor is "the fastest".
 
>"What are you paying for?"

Their paying an "independent tester" to show that the Intel CPU is faster than the AMD CPU.

Do you think they'd be doing this if the testing found that AMD was faster?
 


https://www.youtube.com/watch?v=YeXtTYAPzXU&t=2s

Here's 1000's of comments from the AMD camp that say Xeon benchmarks are relevant for gaming so long as the comparison favors AMD. Completely hypocritical. AMD backers couldn't care less about the fairness of benchmarks. All they care about is AMD looking better than Intel. Unfair benchmark setups are completely OK with them, so long as AMD is the one with the unrealistic advantage.
 


Classic talking out of both sides of your mouth AMD argument. Basically every post complains about the HSF. Almost all of them saying the comparison is unfair because Intel got an aftermarket cooler while AMD was stuck with it's stock cooler. A significant portions of those posts then noting the additional cost of buying a HSF for the Intel side, that doesn't need to be done for AMD. You have to pick one argument or the other, you can't use both. What's the point of the included cooler AMD ships if it isn't good enough to maintain boost clocks? I'd like to know why paying for a cooler you aren't going to use because it hurts performance so much is such a great feature. How about offering CPU's for $20-25 less that don't include a HSF I don't want in the first place?

And let's be realistic. How many enthusiasts buy a new cooler every CPU upgrade? I have a water cooling loop on my system right now. When I upgrade, I'm not going to buy another water cooling loop. I'll migrate that to the new system, meaning I'll pay $0 for a cooler regardless of whether I go AMD or Intel. My current system will be given away to someone and I'll just slap an extra HSF I have lying around on it. If I didn't have one, they could buy it themselves, they're getting a free system after all.
 

Huh?! Are you talking about the same bench as everybody else?It shows intel 20-30% faster in a few tests and even up to 48% in CSgo.

 

I don't think the "best deals" articles have ever been about best performance per dollar or even endorsing the products listed in the article. It's more about listing cases where a product can be found at a significant discount compared to past prices or current prices at other retailers. Whether or not the product is a good buy at the 'deal' price is a different question IMO.
 
Not really an accurate comparison because the amd and Intel releases keep coming out 6 or 7 months apart. 8th gen releases and 7 months later comes ryzen 2, 6 months after that comes 9th gen, another 5 to 7 months from now comes ryzen 3..

Ryzen 2 answered 8th gen, 9th gen answered ryzen 2 and ryzen 3 answers 9th gen, it's really not an accurate comparison. Intel would probably still take the lead even if the release dates were closer, but it's just not worth the extra cost. 50%+ more price for roughly 10% or less performance in avg
 


CSGO:
442.4fps = 9900k
298.1fps = 2700x
265fps = fastest monitor money can buy
0 = real-world difference between CPU fps performance in that test

...numbers alone do not provide context. they are essentially equal in CSGO, so the data result is an outlier & should only be included for the sake of bench thoroughness to find anomalies. the 48% gap is valid, and should be noted, but it's being used in this case to skew the final/combined grading curve in a way that does not represent real-world experience.
 

This is not a benchmark of real-world experience,this is a CPU benchmark

and one CPU is 48% faster than the other.
Just because the difference get's hidden today behind weak GPU performance (or monitor restrictions) doesn't make this difference any less important or any less real,in 4-5years from now the 48% will be the norm instead of an outlier.
 


the entire purpose of benching is to show real-world perf expectations that the end-user will experience, otherwise we could just pull numbers out of a hat that nobody would be interested in. so, like i said; "the 48% gap is valid, and should be noted" ...as it does show a potential gain in certain circumstances.

...if AMD had some sort of tech that could read/write data to/from storage at 10PB/ns, then it'd def be worth noting. but until the rest of the ecosystem catches up to take adv of it, then including it as a win against Intel would be misleading at best. now, if the ecosystem had that capability within it's reach in a relatively short time, then the tech gains noteworthiness. it's academic at this point, just like 440fps.
 
@terrylaze, the difference is not important due to hardware not existing to capitalize on it. In 4 to 5 years when 500hz display exist and bionic eyeballs can notice the difference between 490 and 500fps then sure it matters a lot. That would be a realistic scenario in which there is a real benefit.

Sure the score is higher, but what benefit is there? What does someone gain from this other than saying I can run cs go at 500fps.

 


Yes, lets be realistic. For LESS than the Intel chip, you can get the exact same high end cooler (or go for a good AiO if you prefer) and still be cheaper than Intel. Allowing Intel to have the better cooler is like taking your car, telling you you cannot upgrade from stock, and then pit your car against one that has a supercharger on it. No Contest, (both using pro drivers) The other car will win. by a larger margin than it would otherwise. THAT is the crux of the situation.
 
I don't know how they "test" but if they use software its going to give different results based on how well they made the hard drive system. Typically its a cheap controller that requires a driver and takes up CPU processing. So if they test with the same hard drive controller card that doesn't lean on the cpu, I can see that would be more valid test.
 


Let's continue to be realistic. If your primary goal is the best performance in games, it doesn't matter how you optimize the AMD CPU's. They aren't going to beat the Intel chips. You don't need to buy a 9900k to beat AMD's best, there are multiple cheaper Intel chips in the price range of the 2700x that will beat it. The 9700k is over $100 cheaper than the 9900k and should perform exactly the same as the 9900k in gaming. There is really no reason to buy a 9990k over a 9700k for gaming.

If the rumored 5.3-5.5Ghz for the STIM i9's are true with overclocking, then AMD will not be close. The HSF you use for the 2700x won't matter in the slightest.

Your car analogy makes no sense. Stick to CPU's.
 
Intel makes the best CPUs that money can buy. AMD offers the best value in terms of money vs. performance.

Intel signed off on the report, so it was well aware of what it contained, including how the tests were conducted.
 


yes i am , but seriously who cares if you go above 140+fps even 60fps for that matter? who cares if you do 1000fps on a game. yes the 9900k is "faster" but for that price? and just look at the difference on the other tests it doesnt justify its price tag. also mind you this is a "9th" gen processor from intel versus AMDs answer to the 8-series.

the intel fanboyism in this thread is mind boggling. accept the fact that intel played dirty again to make its self look big, its pathetic and desperate.


Now, I loved intel. for my whole life ive only owned intel processors but this is just utter BS and its sad they have to stoop this low to save themselves its really really sad .

now imagine what the 7nm AMD chips could do.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom