• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

AMD Ryzen 7 2700X 3.7 GHz

Joss said:
What a wonderful way to show a chart! When you hover a certain CPU score it shows 100% with the others relative to it :)
That site is on my bookmarks now ;)
Click to expand...

As far as i know, Computer Base has been doing that in their CPU / GPU reviews for quite some time: dunno of any others that do the same.
 
Slightly off topic

Ryzen second generation Pro coming 2H

Is that Q4???
 
intrepid3d said:
Interesting, TechSpot made a clock for clock comparison, it seems Ryzen 2 has about a 7% gain in IPC in games like BF1, about 3% in cinebench.

That's as much as the whole gain from the 1800X to the 2700X in the same game in this review.



View attachment 100045
Click to expand...

You really should stop Intel shilling, every review site has similar but ultimately different hardware for testing (Silicon lottery?, different Motherboards?, different memory/timings/frequencies) , if your expecting every review site to line up 100% with their results your truly lost and at this point just being a troll. If you don't like the reviews here go some where else that has more 'respectable' numbers and views, it's not like your constant arguing is going to change the review or the results. Give it a rest already.

Thanks @Wizzard for the time and effort you put into the review.:lovetpu:
 
This looks really good. Still some considerable gaps in gaming scenarios but regardless, for 8c/16t this is another great step forward. Intel still wins on clocks though, so for gaming Ryzen still isn't the optimal choice.
 
Now this is interesting. Gaming performance is to be expected... especially on 1080p & 1440p, so no surprises there. Price wise it's very competitive, should be able to shake up Intel a little bit.
 
I wonder what's the percentage of people playing high refresh rates (144hz and above) because I suspect it's diminutive.
And even those who buy 144hz monitors, do they all really demand 144fps in every possible scenario? because I suspect many are happy to play above 100hz.
I find the 2700X to be an excellent solution for the majority of demanding scenarios, gaming only included.
 
What this thread leads me to conclude is not so much about the performance of ryzen 2,that was pretty much to be expected, as it is about the fact that people need to learn how to interpret the numbers.
I said early in the thread that "7600k-like performance is mediocre". It is. 2700X should be at least level with 8600K when it comes to pefromance numbers. But saying that i3 is as good of a gaming processor as Ryzen is simply not true, Ryzen wins is a landslide overall. Take 10 games and let's say that in the end 7600K and 2700X are withing 1-2% of each other. But anyone who has experienced a cpu bottleneck with a fast GPU and high refresh monitor will tell you that 55 fps on a CPU that's using 50-60% across the cores is gonna feel much smoother than 55 fps on a 4c/4t CPU that's getting +90% on every core, or even 70 fps on a 4c/4t CPU that's getting +90% on every core. It just is, I experienced that myself and tested it many times with 4790K HT on vs HT off. The only exception is the broadwell-c family with l4 cache which runs smooth, without stutter or hiccups, even when the cpu is being hammered. The only games that 8350K will feel smoother are those which use 1-2 threads heavily but don't put much usage on other 2 cores. However, they're like 2 out of 10 these days. Far Cry 5 would be an example, they're still using Dunia engine. I think Kingdom Come: Deliverance is another one. The other ones are either decent or very good at multithreading.I bet you anything that if you could play them side by side, that 160 fps in BF1 would feel smoother than 170 fps in same game with 8350K,which would get occasional stutter and hiccups when the cores are heavily loaded. Why is 8350K getting a higher number then - cause faster single core is not equivalent to more threads, and vice versa. What is unquestionably true is that in games that run multithreading well, a cpu that is 50% loaded will not get the fratetime spikes and stutter that a +90% loaded cpu will get, even if the latter produces twice as many fps.
So does ryzen 2700x have mediocre gaming performance ? Yes, it does, cause 8600K is faster and cheaper. Is 8350K as good of a gaming CPU- hell no.
 
Last edited:
intrepid3d said:
by reducing image quality settings one is reducing the workload on the CPU
Click to expand...
No. You literally learned nothing here.

I am gonna have to -1 you for again spreading misinformation about something we already explained to you.

Open any game, look at you cpu usage. Then drop the resolution down to 720p. Do you see the cpu usage go up or down ?

Report back.
 
cucker tarlson said:
No. You literally learned nothing here.

I am gonna have to -1 you for again spreading misinformation about something we already explained to you.
Click to expand...

Where did you explain it? i must have missed that and i'm genuinely interested.

PS: down vote all you like the rep system is utterly meaningless to me. If anything i think its slightly amusing as those things on any forum, especially reddit get used by immature people hammering their mouse button on people who say things they simply don't like or just disagree with them.

In fact what makes you think anyone should worry about the like or dislike buttons? They just make people say whats popular in order not to lose those shiny like points, go head, knock yourself out with it and down vote all you like.
 
intrepid3d said:
Where did you explain it? i must have missed that and i'm genuinely interested.
Click to expand...
https://www.techpowerup.com/forums/threads/amd-ryzen-7-2700x-3-7-ghz.243209/page-2#post-3830814

You should be pretty embarassed by now. No wonder you missed it, you've been posting non-stop since the review came out. Take a shower :D



1440p low quality, 56% usage on CPU, 86% usage on GPU

0AR7nWp.jpg


1440p Ultra, 38% on the CPU but 97% on the GPU

KMXnrfh.jpg
 
Last edited:
intrepid3d said:
Yes, and i stand by that.

I understand and agree with 720P testing but by reducing image quality settings one is reducing the workload on the CPU. i don't think i need to explain it beyond that, if you think about it for a minute it should make sense to you.
Click to expand...
Your techspot link uses medium quality
 
cucker tarlson said:
Open any game, look at you cpu usage. Then drop the resolution down to 720p. Do you see the cpu usage go up or down ?

Report back.
Click to expand...

That's pseudo logic, by turning the resolution down the GPU is rendering faster which makes the CPU work harder to keep up.
However, games are far more complex than that, Graphics quality settings are there to reduce the workload on the whole system, including the CPU, in games the CPU runs Streamed shading and lighting calculations, post processing filters, soft and ridged physics math.... a whole bunch of things and by reducing graphics settings you are reducing or in some cases turning off those things and in that way the CPU is doing less work.
Usually by reducing graphics quality settings you are reducing the need for higher core count CPU's given that a lot of these things like Physics are very parallelized workloads so you can make a 2 or 4 core CPU look just as fast as a 6 or 8 core where in fact when such things are turned on they overwhelm small CPU's with less cores.
Observe this video, why do you think when looking at the sky the 7600K is faster than the Ryzen 1600, but when looking at the complex scene the Ryzen 1600 is twice as fast, because in the complex scene there is vastly more work for the CPU to do and the 4 thread Intel is only half as fast as the 12 thread Ryzen, even with much lower clock speeds.
You turn the graphics settings down and the scene will be far less complex allowing the 7600K to catch up again, its why i'm always suspicious of reviews that have 4 thread CPU's and 12 thread CPU's from the same vendor clumped together..

Edit, this forum will not allow time stamp youtube :( go to 5m10s

 
You'd have to have a cpu and gpu load figures in that video to prove that. They're not there, so either produce your own resluts with usage included like I did or just give us a rest. You're basically saying that a higher cpu usage numer means it has less work to do. If what you're saying was true, then a faster CPU would make a difference at 4K. It never does.
 
cucker tarlson said:
You'd have to have a cpu and gpu load figures in that video to prove that. They're not there, so either produce your own resluts with usage included like I did or just give us a rest. You're basically saying that a higher cpu usage numer means it has less work to do.
Click to expand...

Wow... you didn't even whatch it, if you had you would realize this ^^^^ is a nonsense. did you go to time stamp 5m10s and watch it from there?
 
intrepid3d said:
Wow... you didn't even whatch it, if you had you would realize this ^^^^ is a nonsense. did you go to time stamp 5m10s and watch it from there?
Click to expand...
Of course I didn't watch it. You just said "when looking at the sky.." and then posted a 17 minute video with no reference to the scene you were talking about.

I've wasted enought time talking to a troll.
 
cucker tarlson said:
Of course I didn't watch it. You just said "when looking at the sky.." and then posted a 17 minute video with no reference to the scene you were talking about.
Click to expand...

This forum removes Youtube time stamps, so you have to go there yourself... 5m10s watch 2 minutes or more if you like from that time. again 5m10s
 
All this proves is 4c/4t is too few cores to run crysis since this game is very heavily multithreaded. It has nothing to do with decreasing resolution and how it impacts cpu usage. You're not comparing identical scenes like you do with a 1080p vs 720p tests here, you're comparing sky vs terrain. 7600k renders sky flawlessly, but it can't cope with terrain. If you dropped the resolution in that scene from 1080p to 720p, that 7600K would be even more loaded, that is if it hadn't already hit 100% usage before, but we don't know that since there is no cpu or gpu usage.

I think you're just drawing the wrong conclusion. You think that if 7600k can handle rendering the sky but can't cope with terrain, then increasing image quality will always hammer the cpu. It is not true in 99.9% of the cases. It means some scenes in the game are just easier to render than others,whether it's gpu or cpu that has it easier depends on the scene. This is the conclusion you should have drawn.
 
Last edited:
cucker tarlson said:
All this proves is 4c/4t is too few cores to run crysis since this game is very heavily multithreaded. It has nothing to do with decreasing resolution and how it impacts cpu usage. You're not comparing identical scenes like you do with a 1080p vs 720p tests here, you're comparing sky vs terrain. Think for a second dude !
Click to expand...

You're the one fixated on resolution, i have said it many times already to try and be clear but i'll say it again, reducing the resolution is perfectly valid, i have no problem with it, in fact i agree testing at low resolution should be included in reviews, did you get that into your head now?

Crysis 3 is a 6 year old game, these days it is far from unique, almost every modern title has areas that put a lot of emphasis on more rather than less threads.

They are not difficult to find if you know what to look for, if you understand how game engines work and how they use your CPU, as Digital Foundry do because in the same video they also identified this in Ryse of the Tomb Raider and Assassins Creed, games that are usually seen as very heavily Intel bias, only when you benchmark simple scenes.

Games the world over now have a mixture of simple and complex scenes that in one instance can make a high clocked i3 look better and in another instance can make a low clocked Ryzen 1600 look better.
 
intrepid3d said:
in the same video they also identfied this in Ryse of the Tomb Raider and Assassins Creed, games that are usually see as very heavily Intel bias, only when you benchmark simple scenes.
Click to expand...
Again, this is cause faster IPC and increased number of threads are not equivalent, and when the intel cpu is not getting 90-100% loaded on all threads but has some breathing space it will pull away from amd.

Yes, I got that into my head now, although I actually never undermined 720p testing nor did I say it's invalid. I just caught you pants down saying stuff that's just plain wrong, that's the origin of the discussion. I bid you farewell.
 
cucker tarlson said:
Again, this is cause faster IPC and increased number of threads are not equivalent, and when the intel cpu is not getting 90-100% loaded on all threads but has some breathing space it will pull away from amd.
Click to expand...

So what? how is it the fault of the AMD CPU with its higher thread count that the Intel CPU cannot keep up with it in complex scenes?

Where or how is the justification that Intel should be given the benefit of not being tested in a way that puts full load on their CPU's simply because in such scenario the AMD CPU is clearly faster, how do you reason that?
 
Ugh, takes a lot of explaining, but here's the short version

Where or how is the justification that Intel should be given the benefit of not being tested in a way that puts full load on their CPU's, simply because in such scenario the AMD CPU is clearly faster, how do you reason that?
Click to expand...

1. except all that you wrote is actually true. yes, they do test them under full load scenario, and yes intel's 4 core / 4 thread cpus tend to get hammered hard when the game requires 6 cores or more.

https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,38
https://www.purepc.pl/karty_graficz..._s_creed_origins_problemy_w_egipcie?page=0,12
https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,37
https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,31

2. and this is like the main theme here, you are drawing waaaay too much conclusions just from this one digitalfoundry test, and from one oddball scene in particular.I'll tell you I've ran many cpu tests myself, I've seen huge differences between i5 and i7, but never been able to reproduce such a huge margin, under no circumstances. 7600k is faster than 1600 when the game is not using all threads, and it gets hit hard when the game suddenly requires 12 threads to run smoothly. But that's what I've been trying to get across to you, this is the conclusion you should have arrived at: 4c/4t intel is faster than amd when the particular scene in game doesn't require all of its resources. It gets a heavy performance drop when those resources are running at their limit in other scenes while amd ryzen still has plenty of cpu resources so it doesn't get such a drastical performance drop. Instead this is the conclusion you arrived at: "I would have thought by now people are clued up enough to know that by turning graphics setting down you're reducing the load on CPU's ". Swing and a miss.

I literally ran out of time I should spend talking to you like 20 minutes ago. Take care.
 
Last edited:
cucker tarlson said:
Ugh, takes a lot of explaining, but here's the short version



1. except all that you wrote is actually true. yes, they do test them under full load scenario, and yes intel's 4 core / 4 thread cpus tend to get hammered hard when the game requires 6 cores or more.

https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,38
https://www.purepc.pl/karty_graficz..._s_creed_origins_problemy_w_egipcie?page=0,12
https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,37
https://www.purepc.pl/procesory/tes...e_i3_8350k_prawie_jak_core_i5_7600k?page=0,31

2. and this is like the main theme here, you are drawing wat too much conclusions just from this one digitalfoundry test. yes, 7600k is faster than 1600 when the game is not using all threads, and it gets hit hard when the game suddenly requires 12 threads to run smoothly. But that's what I've been trying to get across to you, this is the conclusion you should have arrived at: 4c/4t intel is faster than amd when the particular scene in game doesn't require all of its resources. It gets a heavy performance drop when those resources are running at their limit in other scenes while amd ryzen still has plenty of cpu resources so it doesn't get such a drastical performance drop. Instead this is the conclusion you arrived at: "I would have thought by now people are clued up enough to know that by turning graphics setting down you're reducing the load on CPU's ".

I literally ran out of time I should spend talking to you like 20 minutes ago. Take care.
Click to expand...

Ok i'll put it a different way, i take it you watched that video now? and you saw the massive frame rate dips on the 7600K, the stuttering that was happening on it?

None of that shows on slides, in fact slides can be very forgiving because taken over a longer period they dilute the performance differences given it is taken from averages over a period of low and high frame rates.

Whats more they don't show you that stuttering, its why i argue slide only reviews these days can be misleading,. because they hide the true gaming experience one can expect from the CPU's raw numbers in text only.

This is Insurgency, an old source engine game, its me, this is me playing this game, this very old game, the GPU is a GTX 1070, the CPU a 4.5Ghz 4690K.
Watch closely what happens to the 4 cores in this , they jump around bouncing off 100% on all 4.... i can tell you it feels horrible, its laggy and stuttery, in some places you can even see that stutter.

Now here's the thing, if i didn't already know about this from owning an Intel 4 core, and as a noob i'm looking to buy a CPU, if i look at w1zzards slides i would be lead to conclude the 4 core Intel i3 is much better than the 1600, the 2600, even the 2700X, when in fact the experience i would actually get, is as it is right now for me, pretty dreadful, i would have been much better off even with the Ryzen 1600.


Its not to have a go at W1zzard, i just hope that as a good reviewer he takes this feedback for what it is.

BTW a lot of the games i play are as bad or worse than that, on this Intel 4 core.
 
Last edited:
You must log in or register to reply here.
Back
Top