What are the undervolting settings and what does it score without it? Best mine does is around 980, but I haven’t tried undervolting.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
My new base 2.2 blows away my i9!
- Thread starter drvelocity
- Start date
- Sort by reaction score
What are the undervolting settings and what does it score without it? Best mine does is around 980, but I haven’t tried undervolting.
The numbers are not as stark as I thought they would be, but I pulled the screen grab from my post patch test, and the second is with undervolting
I thought Apple locked it so you couldn't undervolt? Are you sure it isn't just a placebo effect?
Apple uses unlocked chips and I'll take the placebo effect of having my MBP run faster
. Running Cinebench multiple times, w/o undervolting did not come close to the values I was seeing with undervolting. Likewise the numbers with undervolting were always consistently higher.
Volta has been updated, the prior update locked the undervolt slider, looks like the developer was able to update his app and provide the functionality. I went all the way to -.145. I'm not sure I want to push it any further.
I received my 2.2/32/512/560X this morning. It seems I've won the silicon lottery
No undervolting, No adjustments to fan speed, Ambient room temperature of 24 degrees Celsius and I'm getting 1069 the first two times.
After 10 runs of Cinebench the average score is 1045.
No undervolting, No adjustments to fan speed, Ambient room temperature of 24 degrees Celsius and I'm getting 1069 the first two times.
After 10 runs of Cinebench the average score is 1045.
Attachments
I haven't tried this, but I think Throttlestop may be your answer in Windows. I'll probably rebooting into windows later today.
Scroll up, its already there
https://forums.macrumors.com/threads/my-new-base-2-2-blows-away-my-i9.2129384/page-3#post-26292597
MacOS is basically FreeBDS and if you have a lot of spare time, you can look at the kernel source code here: https://opensource.apple.com/source/xnu/xnu-4570.41.2/
Swap is simply the memory behind the RAM. The OS will use it if its finely tuned algorithms, written over decades by some of the smartest people around, decides that it makes sense to write some of the memory pages to the disk. And there can be a lot of reasons for it.
So unless you are running into quantifiable performance issues, I wouldn't even look at the swap usage.
Scroll up, its already there
https://forums.macrumors.com/threads/my-new-base-2-2-blows-away-my-i9.2129384/page-3#post-26292597
Apologies. I thought these results were with an older version of Volta that didn't support undervolting.
It's surprising to see there are such differences in the scores of the same model. (I know those differences are very very minor and there are lot of factors involved but still interesting to see)
I think you might be over-interpreting it. To me it seems very much like an identical score, within statistical error. Frequencies, power draw, temperatures — its exactly the same. I don't think that the undervolting has worked?
Throttlestop won't even start on this machine, was the first thing I tried out a few days ago. Intel XTU has the voltage part locked as well. I will try again later, keep me posted if you have any luck with Throttlestop.
I hear what your saying, however, judging by the quoted piece below, it seems;
A. Maflynn said undervolting is working now after the update (or did I misinterpret that?)
B. I'm talking about the scores of my factory state computer. I did not modify anything yet. Therefore I came to the conclusion there is in fact a difference, since maflynn already modified things (Volta, Fans) to get to these scores.
If I'm wrong or misinterpreting these scores please let me know. I just want people to get the maximum performance out of their new machines and spread the knowledge
Would love to know how the 2.6 fairs. With the SSD and GPU upgrade from a 2.2 it's an $80 diff.
If I'm wrong or misinterpreting these scores please let me know. I just want people to get the maximum performance out of their new machines and spread the knowledge
I think we might be both confused, so I can't comment on that
My only point is that results before and after undervolting seems basically identical. This is also in the ballpark of what my i9 machine gets (without any modifications).More on topic: in multi-threaded tests like Cinebench, all three CPUs will behave similarly. The MBP is able to dissipate around 50-55Watts in sustained CPU operation (which is very much on par with other machines such as Dell XPS 15" or Razer Blade), which is enough to keep the CPU at around 3.2-3.3Ghz (all cores loaded). Silicone lottery aside, this is within the turbo limit of each of these CPUs and hence I am not surprised about their similar performance. The reason why I am keeping my i9 is single-core boost (I've seen it reach 4.5Ghz, which is a nice additional boost for burst work) as well as 12MB cache.
Why don't we look at single-core scores? I'd expect more difference there.
Thats the thing. From a pure value stand point for someone looking for 512GB the 2.6 is the way to go.
Correct me if my numbers are wrong.
I posted something earlier about Cinebench and how I would take their scores with a grain of salt.
Here it says in this test done by PCMAG that the i9 Alienware gets 1036. Does this mean that it will perform worse than you guys getting 1040-1060 cinebench scores? I don't think so.
I have the both the 2018 2.2 and the i9. I have nothing to be insecure about. We all just need to be more informed than creating an uproar and basing all our computers on a single benchmark application and then you end up with people getting disappointment, angry and all these other emotions which to be honest, these numbers should never be the basis of absolution.
[doublepost=1532699948][/doublepost]
Thank you. As mentioned, I believe these benchmarks are good but it doesn't mean an absolute cause to say THIS is the performance of a machine which is exactly what some people are doing here. Not sure if you have all seen this but this shares some insight on performance. There are some instances that the 2.6 does perform better or equally to the i9.
https://translate.googleusercontent...700208&usg=ALkJrhj_ZniTBw4JPV1BcVE_9Cdx9_Bm7A
Last edited by a moderator:
No I disagree because simply I cannot get close to that number w/o undervolting and I never dip close to the non-undervolting number. That is I have hard evidence that I'm undervolting. You may not choose to believe that, but I think the numbers stand on its own.
I agree that the undervolting numbers are not far off to the non, I thought I'd see a bigger gap, nonetheless I cannot reproduce that higher number UNLESS I undervolt.