Overclock.net - An Overclocking Community - Reply to Topic

Thread: [ZoLKoRn] R7 vs i7 - core for core - clock for clock Reply to Thread
Title:
Message:

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in


  Additional Options
Miscellaneous Options

  Topic Review (Newest First)
04-02-2017 05:01 PM
Raghar
Quote:
Originally Posted by philhalo66 View Post

because if you need to underclock a processor then its unfair. same if they were to underclock ryzen to compare it with a celeron.
What's problem with that? Normal reviewer would set both CPUs to the same frequency to show differences at the same speed, and then he would set it to the same different frequency to show frequency scaling. For Ryzen seting it to lower frequency is vital to remove XFR. For Intel disabling turbo just evades all crazy stuff MB can do on auto, thought a competent reviewer NEVER writes review with only auto tests.
03-30-2017 10:07 PM
Tsumi
Quote:
Originally Posted by philhalo66 View Post

so they underclocked the 7700k? kinda seems unfair to me.

The comparison is essentially useless for desktop parts, but it would be very useful for comparing mobile parts. If you know Ryzen is only about 5% in IPC, then when looking at a laptop, you would know a 2.6 ghz Ryzen laptop would essentially have identical CPU performance to a 2.5 ghz Skylake/Kaby Lake laptop.
03-30-2017 05:42 PM
octiny
Quote:
Originally Posted by philhalo66 View Post

because if you need to underclock a processor then its unfair. same if they were to underclock ryzen to compare it with a celeron.

Which Celeron? Compare it to any Celeron just disable the cores/smt and underclock it. Compare it to the FX line, same thing. It's all relative. Just like comparing Broadwell-E to SK/KB, adjust to same clock speed to see the improvements in IPC. It's not exact science obviously, but it's what people have done for the past 15 years+......until Ryzen got released, and all of a sudden it didn't matter anymore rolleyes.gif.

The whole point of it is to see the architecture improvements. Which for Ryzen is HUGE and bodes well for further optimization and Zen+/2.
03-30-2017 05:26 PM
philhalo66
Quote:
Originally Posted by octiny View Post

How's it unfair? They're essentially comparing the IPC, the only way to do that is to run both at the same clock speeds.

Ryzen is much closer to SL/KB IPC than a lot of people thought.

Looking forward to all the optimization that'll occur over next 12 months.......and even more excited for their X390 platform (if the rumors are indeed true).

because if you need to underclock a processor then its unfair. same if they were to underclock ryzen to compare it with a celeron.
03-30-2017 05:21 PM
octiny
Quote:
Originally Posted by philhalo66 View Post

so they underclocked the 7700k? kinda seems unfair to me.

How's it unfair? They're essentially comparing the IPC, the only way to do that is to run both at the same clock speeds.

Ryzen is much closer to SL/KB IPC than a lot of people thought.

Looking forward to all the optimization that'll occur over next 12 months.......and even more excited for their X390 platform (if the rumors are indeed true).
03-30-2017 05:14 PM
Asmodian
Quote:
Originally Posted by philhalo66 View Post

so they underclocked the 7700k? kinda seems unfair to me.

When comparing clock for clock it is critical to have them running the same clock. smile.gif

It is a very interesting test, not to show which CPU is "better" because there are too many factors, but for academically understanding the architectures.

It looks like Kaby Lake has a bit better IPC than Ryzen while Ryzen's extra cache helps, but only rarely.
03-30-2017 05:04 PM
philhalo66 so they underclocked the 7700k? kinda seems unfair to me.
03-30-2017 04:26 PM
KarathKasun
Quote:
Originally Posted by Quantum Reality View Post

One thing I'd like to point out that doesn't get as much attention as it should is the condition of the electrical power entering a PSU. There have been some weird cases discussed here on OCN where in one case, a particular power outlet seemed to "kill" computers and it was traced down to voltage spikes that the PSU couldn't "damp out" effectively, and so the electrical components simply degraded over time.

Sometimes putting a good UPS unit with a true sine wave output can help, because they are designed specifically to deal well with power spikes.

Multiple failures in different houses, and that is getting into the very far end of outlier cases.
03-28-2017 10:42 AM
Quantum Reality
Quote:
Originally Posted by KarathKasun View Post

Quote:
Originally Posted by Formula383 View Post



I really love the generalizations, How long is long term? 20 years or 2 years? or maybe 5 years? Maybe you have hard time understanding my point? let me explain it for you. The point is cpu's will take a lot more abuse than people give them credit for. A voltage of 1.6v or more on SB-E is crazy and should NOT be run period.

Yet i have run it above that. And have stressed it at times for hours at a time. Albeit not for 24/7 100% load. But how many people that are interested in overclocking use there cpu 100% load 24/7? I have idea almost none. Because anyone with half a brain knows a cpu at full load 24/7 will run a risk of dieing at some point (how long before it dies who can say). Not to mention the absurd amount of wattage overclocking will pull from the wall @ 100% load 24/7 there is just no point in it. You would be better off just running 2 systems and getting more work done faster for the same cost to run. Over clocking is for people who game and or are a power user. And i have news for ya power users dont run there cpu at 100% all day while they work on it. Unless its a business and that is the task for the pc. But there again not for overclocking... 99% of people saying cpu's will die from high voltage have no idea. Its like console guys saying there console is better than a pc at gaming.

More over getting to why i posted this in the first place so you under stand its context, It is reasonable to say people with kabylake cpu's can and will be running them at 5.0ghz for long term use (no maybe not everyone). 1.4v or less is not going to hurt that cpu Intel even says it can run up to 1.52v! on the data sheet... common show me where you cpu dies from 1.4v. Further more even if A cpu died at A voltage it does not mean in any way that is going to happen to everyone else or anyone is at a high risk of it happening.

There are simply too many variables that CAN cause damage to your cpu. Simple if YOU think that voltage is not safe then dont run it. But dont run around telling everyone else in the world that X voltage is FARR to much when you have ABSOLUTELY no experience with it. IF you do then share it. Simple. People running off spouting things they know nothing about is my problem here. Oh i read it on the internet so there for it must be true! LOL. Point and case IF 1.4v on SB would kill a cpu you would most CERTAINLY believe 1.625v would very shortly kill that cpu and you are right it WOULD. But it has not. Simple as that. Yet people insist that they know better than the facts. If you dont know dont say it please. I do not think low voltage is a bad thing. But voltage is REQUIRED to make a cpu stable at a given speed.

Proof? And not just someone saying it. What evidence of this do YOU have?
Ya i'm sure they just put that voltage in there for no reason it means nothing just ignore the data sheet lol.

You know its down right silly. Would you believe my asrock pro4s z170 board has a bios max cpu vcore of exactly 1.52v. Can you belive it? hmmm i wonder.

3 CPU's in the bin that were never out of spec thermally, and were at 1.4-1.45v (SB). All started slowly going downhill after around two years. All were running some variety of tasks 24/7 and I would say they were averaging a load of about 50% over that time period.

5+ NW P4's that started dying within a year from less than a 20% voltage bump, with very low temps (55c or under).

A whole slew of AXP's that lasted maybe 3 years at ~1.85v (not a huge voltage bump for them, 2.0v was doable with water).

2 A64's that died with ~1.6v (also not a huge bump) after ~2 years.

Had an A64X2 that ran much hotter at ~1.55v and is still kicking after 5 years of hard use. Same with my Phenom 9950 rig, lived its whole life at 3.2ghz with relatively high voltage and had zero problems.


Point is, for many chips it takes nowhere near as much voltage to degrade the chip as you think. My 18 years of overclocking points to this being the case. If you can afford to replace your CPU every year or two, pushing high vcore to get high clocks is not a problem. If you are going to hold on to the equipment longer than that or plan to push it into some sort of retirement role (storage/game server, hand me down system, etc.), stick to within 10-20% of the stock voltages.

One thing I'd like to point out that doesn't get as much attention as it should is the condition of the electrical power entering a PSU. There have been some weird cases discussed here on OCN where in one case, a particular power outlet seemed to "kill" computers and it was traced down to voltage spikes that the PSU couldn't "damp out" effectively, and so the electrical components simply degraded over time.

Sometimes putting a good UPS unit with a true sine wave output can help, because they are designed specifically to deal well with power spikes.
03-27-2017 11:51 AM
BiG StroOnZ This is pretty impressive stuff:







It's a step in the right direction, IMO. However, I'd be looking forward to see if with revisions (Zen+) we can get those clockspeeds up and memory speeds higher.
This thread has more than 10 replies. Click here to review the whole thread.

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off