Personal computing discussed

Moderators: mac_h8r1, Nemesis

 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Voltage / frequency curves

Tue Jul 18, 2017 12:46 am

No reason other than "I was curious". This is not perfectly professional (as is obvious from Vega's inclusion); I took pains to make the graph true to the original data, but the data sources are not uniform. My two were captured by running prime95, adjusting max CPU frequency in OS power management tools, and watching Vcore (at stock settings, no OC involved). Vega's also appears to be the stock power management's curve, though of course it doesn't quite fit here (and I doubled its frequencies to put it on a more comparable part of the graph). The Zen curve is from The Stilt's measurements here. I don't know how he captured the data, but he clearly knows what he's doing.

More data would be great! I wish I had more hardware around to test. If you post it, I'll add it to the graph. Sparse datasets (to make it quicker to measure) are no problem, especially at low voltages where the line should be straight anyway.

Here's the original image - the one below is updated.

Image

.... And of course I've got a pile of unanswered (sometimes unanswerable) questions to go along with it:

What parts of this does chip-to-chip variance mostly affect? I thought thresholds were pretty consistent on a given process (not that I'd know), and the two Intel chips here both look to be on threshold at 800 MHz (voltage drifts oddly at that speed and no other, possibly due to temperature). The slope of the line (where straight) must be pretty consistent if it's precisely the same across two chips that different. Zen and Vega both cornering at 1000mV suggests that point may also be fairly constant. That would mainly leave the shape of the line after the corner.

Extrapolating my G3258's line out with no corner results in 4.8 at 1.36, which is much closer to reality than I would have guessed (for Haswell in general). At least some Haswell chips must either corner very late or not do much of anything after the corner. Mine isn't so golden - it used to do 4.3 at 1.33 and has degraded substantially since. That's still a lot better than it looks like Skylake should be able to do by that line. At that rate, a 6700K would take nearly 1.5V just to turbo. This i3-6100 does seem to corner strangely low; 2.2 is actually the last sample on the straight line I got. (Windows ramps it up mostly in 200 MHz increments, but jumped 2.2 -> 2.5, which seems like it might be hiding a spike.)

How do Broadwell and Kaby compare, and how will Cannonlake? It feels like Broadwell and Cannon have similar things going on, but I'm curious what that thing looks like on a graph like this. Does Intel still use different process variants for mobile and desktop? If so, how does their mobile process differ?

Zen and Vega have fairly similar shapes, but for a given voltage bump, Vega gains a lot more speed (as a percentage). That seems odd. Could it maybe be something about transmission limitations in the metal layers versus limitations in the transistors themselves? On that line of thought, does Skylake-X have anything similar going on?

GloFo's 7LP is purported to be rather HP. How much of this wondrous low voltage performance we see Zen turning in will it be able to maintain?

14LPP's marketing slides say 800mV nominal, 945mV overdrive. This doesn't at first glance appear to have much bearing on Zen and Vega's characteristics. Does it actually? If so, can we learn anything from seeing that the equivalent marketing slide for 7LP says 750mV nominal and 850mV overdrive?

I've so far been operating under the assumption that chip wear and power use are affected by voltage changes in about the same way regardless of what the starting voltage is. Is that true? The corner in these lines seems like it could have some other discontinuities associated with it.
Last edited by synthtel2 on Fri Oct 13, 2017 2:38 am, edited 5 times in total.
 
DPete27
Grand Gerbil Poohbah
Posts: 3180
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Voltage / frequency curves

Tue Jul 18, 2017 11:11 am

I started this thread a while back trying to get people to share their stock and custom voltage/frequency curves for their GPUs. Not sure how much of that is helpful to your cause.
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Tue Jul 18, 2017 3:12 pm

Yours is the only one I saw there, and OC data is a bit weirder to start with, but that is interesting. Most notably, yours matches the 14LPP slide by cornering a lot closer to 945 than 1000. Is the extra flattening at low power an artifact of that thing where core voltage is always above VRAM voltage?
 
DPete27
Grand Gerbil Poohbah
Posts: 3180
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Voltage / frequency curves

Wed Jul 19, 2017 3:33 pm

The voltage definitely increases exponentially past 1275MHz. Below that point, yes, the voltage is so close to the 900mV floor established by my VRAM that I just curved it out nicely.
Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Wed Aug 02, 2017 1:05 am

Updated with the R7 1700 I just got. I can't get cpufreq to give me much granularity on that, unfortunately. I've got three modes - 1.55 = 830mV, 2.7 = 910mV, and "3.0" = 1060mV. The so-called 3.0 mode is actually more like 3.15-3.2 AFAICT. Single-core loading gets it to about 3.6 and 1310mV in the testing I've done. I've never seen it reach 3.7. When idle (background load only), it can be anywhere from 670mV to 1310mV (it's very turbo-happy).

:-?
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Thu Aug 03, 2017 8:05 pm

synthtel2 wrote:
Updated with the R7 1700 I just got. I can't get cpufreq to give me much granularity on that, unfortunately. I've got three modes - 1.55 = 830mV, 2.7 = 910mV, and "3.0" = 1060mV. The so-called 3.0 mode is actually more like 3.15-3.2 AFAICT. Single-core loading gets it to about 3.6 and 1310mV in the testing I've done. I've never seen it reach 3.7. When idle (background load only), it can be anywhere from 670mV to 1310mV (it's very turbo-happy).

:-?

At stock, 1.31 V seems to be really high for 3.7 GHz. Then again, I think the current Zen cores aren't really intended for high clocks...

That does look rather high compared to my i5-4590 that does about 1.08V for 3.7 GHz (1.04 for 3.5).
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Thu Aug 03, 2017 9:13 pm

I think most of the Zen OC data I've seen is 1.2-1.3V for 3.7 GHz, so accounting for the extra margin the factory puts on it, 1.31 for 3.6 seems high-ish but close enough to the normal range. My data versus The Stilt's data definitely looks weird at low power, though. That mine goes as low as 670mV when actually idle makes me think some specific part wants more voltage.

Yours is probably more representative of Intel high-performance bins. The two I've got here are both obviously not showing Intel at their best. If you'd be interested in capturing voltage at 900 MHz, ~2.5 GHz, and ~3.0 GHz, I'd definitely like to know how that line trends.
 
bfg-9000
Gerbil First Class
Posts: 184
Joined: Tue Mar 01, 2016 9:17 pm

Re: Voltage / frequency curves

Thu Aug 03, 2017 9:21 pm

Does anyone else find 1.4v on 14nm disturbing? Back at 45nm we had a never-exceed voltage of 1.45v, and at 130nm (nearly 10x larger) we had Sudden Northwood Death Syndrome above 1.7v...
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Thu Aug 03, 2017 10:08 pm

synthtel2 wrote:
I think most of the Zen OC data I've seen is 1.2-1.3V for 3.7 GHz, so accounting for the extra margin the factory puts on it, 1.31 for 3.6 seems high-ish but close enough to the normal range. My data versus The Stilt's data definitely looks weird at low power, though. That mine goes as low as 670mV when actually idle makes me think some specific part wants more voltage.

Yours is probably more representative of Intel high-performance bins. The two I've got here are both obviously not showing Intel at their best. If you'd be interested in capturing voltage at 900 MHz, ~2.5 GHz, and ~3.0 GHz, I'd definitely like to know how that line trends.


I'll try to see what I can do. I think I notice the third core appears to need slightly less voltage than the other cores; I think I'll provide its numbers separately.
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Thu Aug 03, 2017 11:58 pm

bfg-9000 wrote:
Does anyone else find 1.4v on 14nm disturbing? Back at 45nm we had a never-exceed voltage of 1.45v, and at 130nm (nearly 10x larger) we had Sudden Northwood Death Syndrome above 1.7v...

Verily. Here's the thread I made about it in context of Skylake. Skylake at least (I don't know much about Kaby) seems to have a whole lot of variance. Some dies don't take much, and some are like my friend's 6700K or that 6100 in the graph up there. That still worries me more than Zen, because at least Zen only does 1.4+ on a couple cores at a time and under substantial thermal constraints. Some Skylake chips, AFAICT, will happily push 1.36V on all cores at 95C all day.

Noinoi wrote:
I'll try to see what I can do.

Thanks!
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Fri Aug 04, 2017 1:38 am

synthtel2 wrote:
Noinoi wrote:
I'll try to see what I can do.

Thanks!

Cool! Using motherboard automatic voltage, right?

(I'm actually considering undervolting the CPU slightly if I can get away with it after this test.)
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Fri Aug 04, 2017 1:50 am

Yeah, I've just been using stock curves. OC/UV curves are probably more accurate, but figuring out the limits of stability at each frequency is way too much work. In Windows, I just used power options -> advanced plan settings (or whatever it was called, not on Windows to check) to adjust. That's only annoying because it's a percent instead of an absolute speed and pressing enter closes the window instead of making the new value take effect, necessitating mouse use.
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Fri Aug 04, 2017 2:17 am

synthtel2 wrote:
Yeah, I've just been using stock curves. OC/UV curves are probably more accurate, but figuring out the limits of stability at each frequency is way too much work. In Windows, I just used power options -> advanced plan settings (or whatever it was called, not on Windows to check) to adjust. That's only annoying because it's a percent instead of an absolute speed and pressing enter closes the window instead of making the new value take effect, necessitating mouse use.


Funny. At some point with Windows 10, I could just directly specify a target frequency with the same settings. Maybe it's a Creator's Update thing.

Saves me the hassle of rebooting or using XTU though.

Edit: I've gotten some numbers. Ignore the division by 0 numbers; restarting to get the missing numbers is probably too much effort for little gain in accuracy. It seems like the Windows 10's method of limiting processor frequency isn't fine enough to accurately step by 100 MHz intervals, so I had to reboot for the official base clock, 900, and 2500 MHz.

 MHz   Core1   Core2   Core3   Core4   Average
 800   0.702   0.707   0.705   0.703   0.704
 900   0.717   0.722   0.720   0.723   0.721
1000   0.727   0.727   0.730   0.727   0.728
1100   0.741   0.746   0.744   0.742   0.743
1200   0.751   0.751   0.749   0.752   0.751
1300   0.761   0.761   0.759   0.761   0.761
1400   0.776   0.776   0.779   0.776   0.777
1500   0.785   0.781   0.784   0.786   0.784
1600               #DIV/0!
1700   0.805   0.805   0.808   0.805   0.806
1800               #DIV/0!
1900   0.829   0.830   0.828   0.830   0.829
2000   0.839   0.840   0.843   0.839   0.840
2100               #DIV/0!
2200   0.863   0.859   0.862   0.864   0.862
2300               #DIV/0!
2400   0.883   0.884   0.887   0.883   0.884
2500   0.897   0.899   0.897   0.898   0.898
2600   0.907   0.904   0.907   0.907   0.906
2700               #DIV/0!
2800   0.927   0.928   0.931   0.932   0.930
2900   0.946   0.943   0.941   0.942   0.943
3000   0.956   0.953   0.956   0.956   0.955
3100   0.966   0.963   0.966   0.966   0.965
3200               #DIV/0!
3300   0.990   0.988   0.990   0.990   0.990
3400               #DIV/0!
3500   1.034   1.032   1.034   1.034   1.034
3600               #DIV/0!
3700   1.078   1.076   1.078   1.078   1.078
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
Chrispy_
Gold subscriber
Maximum Gerbil
Posts: 4099
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Voltage / frequency curves

Fri Aug 04, 2017 6:13 am

I graphed the voltage/frequency curve for a 290X about 18 months ago to give you 28nm GCN data if you want to look for it.

No idea where it is but I'm pretty sure you'll get it if you search by me as a thread starter - I don't start that many.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
DPete27
Grand Gerbil Poohbah
Posts: 3180
Joined: Wed Jan 26, 2011 12:50 pm
Location: Wisconsin, USA

Re: Voltage / frequency curves

Fri Aug 04, 2017 8:53 am

Main: i5-3570K, ASRock Z77 Pro4-M, MSI RX480 8G, 500GB Crucial BX100, 2 TB Samsung EcoGreen F4, 16GB 1600MHz G.Skill @1.25V, EVGA 550-G2, Silverstone PS07B
HTPC: A8-5600K, MSI FM2-A75IA-E53, 4TB Seagate SSHD, 8GB 1866MHz G.Skill, Crosley D-25 Case Mod
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Fri Aug 04, 2017 3:17 pm

@Noinoi, I updated the graph. Thanks again, that looks much more like Haswell is probably supposed to.

The i3-6100 was on Win7, so there's lots of opportunity for software differences. The same thing applies with the gaps, though.

@Chrispy_ (and thanks DPete27), I don't see any voltages there. It is still an interesting read on the power limits, though.
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Fri Aug 04, 2017 7:10 pm

synthtel2 wrote:
@Noinoi, I updated the graph. Thanks again, that looks much more like Haswell is probably supposed to.

The i3-6100 was on Win7, so there's lots of opportunity for software differences. The same thing applies with the gaps, though.


Alright!

Now I'm beginning to wonder: if we had the data for a Celeron and an i7, it might make the voltage binning differences more obvious. It seems like the Pentium and i3 were of a worse voltage characteristic; a hypothetical i7-4790 or 4790K, or similar Intel CPU, might do as well or better than the i5, unless HT makes it need more.

Just my baseless analysis, though.

Looking at the graph, it does look like Zen corners early and sharply, and probably explains the voltage/MHz wall for overclockers. The wall might come much later for Haswell CPUs and other Intel CPUs, but I don't have an unlocked CPU to test.
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Sat Aug 05, 2017 11:20 pm

4790Ks do need a bit of a lower/later curve than yours has, I think, but not by too much. If yours stayed at that last slope until 4.4 (which it probably doesn't), it would do a 4790K's 4.4 turbo at 1.23V.

Comparing the G3258 and 4590 lines, it looks like at <1V it's pretty close to just a multiplier on frequency. Seems like a probable enough effect for a slow path in a bad chip. One thing that means is that the 800 MHz idle is actually kind of arbitrary, and doesn't match up with threshold voltage half as well as I thought.

Now I'm most curious about the shape for laptop chips. Even if they don't use a different process variant anymore (which seems likely), the binning that makes for a good desktop chip like a 4790K may not be the same as for a top-tier mobile chip.

With enough AMD lines around, it does look like the corner tends to be fairly close to GloFo's 945mV overdrive spec, and some of the inconsistency may be due to manufacturer curves not being right on the line of stability. (Mine probably doesn't show the corner itself due to low resolution.) Between those two Haswell chips, I think it's safe to say there is some variance, though. If overdrive does roughly equal the corner, I think that's enough data (the stuff on the graph + GloFo's overdrive spec + the voltage wall having become more of an OC issue since Sandy Bridge) to say smaller processes do trend towards lower corners. If GloFo really does intend for 7LP to corner around 850mV, I have to wonder at what voltage they think it'll get to 5 GHz.
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Sat Aug 05, 2017 11:56 pm

synthtel2 wrote:
4790Ks do need a bit of a lower/later curve than yours has, I think, but not by too much. If yours stayed at that last slope until 4.4 (which it probably doesn't), it would do a 4790K's 4.4 turbo at 1.23V.


I suppose I might be able to pull numbers from my i7-4700HQ, too, though I suspect speeds beyond 2.4 GHz might be a bit finicky due to my laptop really only having enough cooling to sustain the base clock indefinitely for anything involving AVX burn (Handbrake averages 2.9 GHz; non-AVX loads seem to peg it at 3.1)

Going to install XTU and see what I can do from there.

A sample size of 1 isn't exactly the most scientific, though; hopefully we have other Intel CPU owners here to chime in, if it'd help.
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Sun Aug 06, 2017 1:19 am

Larger sample sizes would definitely be nice, if anyone else is interested. :wink:

Hmm, I was going to suggest a non-AVX load for less heat generation, as I was only using prime95 because it was handy, but now that you mention it Haswells that support AVX (not the G3258) can run different voltages when it's in use. It probably isn't a huge effect on the shape, though, and a lot of Haswells don't seem to have that issue at all (like my 4690K back when it worked). I'll label your 4590 as an AVX load next time I update the graph - it's probably not critical whether it's on or off for anything here, but still good to track.

Now that you mention it, I should see what non-vector loads require from my new CPU at 1.55 GHz. Since true idle is 670, maybe 830 is just some kind of vector thing?
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Sun Aug 06, 2017 2:40 am

synthtel2 wrote:
Hmm, I was going to suggest a non-AVX load for less heat generation, as I was only using prime95 because it was handy, but now that you mention it Haswells that support AVX (not the G3258) can run different voltages when it's in use. It probably isn't a huge effect on the shape, though, and a lot of Haswells don't seem to have that issue at all (like my 4690K back when it worked). I'll label your 4590 as an AVX load next time I update the graph - it's probably not critical whether it's on or off for anything here, but still good to track.


What do you have in mind, out of curiosity?

Anyway, here's some numbers for my 4700HQ. As predicted, anything beyond 2.4 GHz is extremely hot and won't have much time before it throttles back to 2.4 GHz - I figured that I might as well only pull numbers for 3.0. Even the 3.1 GHz stock all-core turbo is too much and immediately resulted in throttling (apparently I'm running into both thermal and current throttling)

 MHz   Core1   Core2   Core3   Core4   Average
 800   0.675   0.678   0.674   0.677   0.676
 900   0.689   0.692   0.684   0.691   0.689
1000   0.703   0.707   0.698   0.706   0.704
1100   0.718   0.716   0.713   0.720   0.717
1200   0.732   0.731   0.727   0.730   0.730
1300   0.742   0.745   0.742   0.744   0.743
1400   0.757   0.760   0.757   0.758   0.758
1500   0.771   0.774   0.766   0.773   0.771
1600   0.786   0.784   0.781   0.782   0.783
1700   0.795   0.798   0.796   0.797   0.797
1800   0.810   0.808   0.805   0.811   0.809
1900   0.824   0.822   0.820   0.825   0.823
2000   0.839   0.836   0.834   0.840   0.837
2100   0.848   0.851   0.849   0.849   0.849
2200   0.863   0.860   0.859   0.864   0.862
2300   0.877   0.875   0.873   0.878   0.876
2400   0.891   0.889   0.893   0.892   0.891
2500   0.920   0.918   0.917   0.921   0.919
2600               #DIV/0!
2700               #DIV/0!
2800               #DIV/0!
2900               #DIV/0!
3000   1.026   1.053   1.024   1.050   1.038


Interesting thing: this sample appears to be worse than my i5 at the high end, but noticeably lower at the low end.
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
Chrispy_
Gold subscriber
Maximum Gerbil
Posts: 4099
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: Voltage / frequency curves

Sun Aug 06, 2017 6:13 am

synthtel2 wrote:
@Noinoi, I updated the graph. Thanks again, that looks much more like Haswell is probably supposed to.

The i3-6100 was on Win7, so there's lots of opportunity for software differences. The same thing applies with the gaps, though.

@Chrispy_ (and thanks DPete27), I don't see any voltages there. It is still an interesting read on the power limits, though.


My bad, I did a lot of voltage trialling for my Polaris undervolting work; Forgot that when I was graphing the 290X it was power limit rather than voltage control.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Sun Aug 06, 2017 1:31 pm

Nice! Graph is updated. Yeah, that's pretty different. Considering the corner is probably actually somewhere after 2.4 (resolution issues), that's probably a more GloFo-like corner steepness, and that really does look like a lower threshold (even if it isn't on threshold at 800 MHz). As of Haswell, Intel does still have a different mobile process?

Maybe encoding something or gaming would be good for decreasing heat load? Prime95 on only one core at a time (like I did to get the full turbo on my 1700) could also be useful, and shouldn't shift any voltages by much.

No worries Chrispy_, still a good read.
 
Noinoi
Gerbil Team Leader
Posts: 228
Joined: Fri Jun 26, 2015 11:31 pm
Location: Somewhere, anywhere!

Re: Voltage / frequency curves

Sun Aug 06, 2017 7:20 pm

synthtel2 wrote:
Nice! Graph is updated. Yeah, that's pretty different. Considering the corner is probably actually somewhere after 2.4 (resolution issues), that's probably a more GloFo-like corner steepness, and that really does look like a lower threshold (even if it isn't on threshold at 800 MHz). As of Haswell, Intel does still have a different mobile process?

Maybe encoding something or gaming would be good for decreasing heat load? Prime95 on only one core at a time (like I did to get the full turbo on my 1700) could also be useful, and shouldn't shift any voltages by much.


That might work; though I suspect it'd be really hard to get gaming numbers (as GPU load-caused heat would be its own can of worms); might opt for video encoding instead and see if I can manage to hit 3.1-3.4 GHz. I might run into current throttling, though.

As you've mentioned resolution issues, I think I'll try to collect even more data points for the missing range, though it does feel like the corner was rather obvious due to the massive jump for 2.5 GHz.

EDIT: I've gotten the full set of numbers for the fourth core for Prime95. It should work mostly for your purposes. I couldn't get it above 3.0 GHz due to current limit throttling - no way around it, I'm afraid, with my laptop.

For the record, using the Intel XTU stress test lets me clock the CPU to the true maximum of 3.6 GHz, but the voltages are lower across the board - 0.97V versus 1.05V at 3.0 GHz, as an example, so it's no longer directly comparable.

 MHz Volts
 800 0.677
 900 0.691
1000 0.706
1100 0.720
1200 0.730
1300 0.744
1400 0.758
1500 0.773
1600 0.782
1700 0.797
1800 0.811
1900 0.825
2000 0.840
2100 0.849
2200 0.864
2300 0.878
2400 0.892
2500 0.921
2600 0.950
2700 0.974
2800 0.998
2900 1.021
3000 1.050


It appears the 1V point for the 4700HQ is at 2.8 GHz; compare to the i5-4590's 3.3 GHz, it is noted that the 1V point on the laptop isn't the base clock. Make of that what you will.
i5-4590 | Kingston 2x8GB | Asus Strix GTX 970 | Asus Z97-Pro Gamer | Kingston Fury 240GB + WD Black 2TB + Blue 2TB | Win 10 FCU
 
synthtel2
Gold subscriber
Gerbil Elite
Topic Author
Posts: 608
Joined: Mon Nov 16, 2015 10:30 am

Re: Voltage / frequency curves

Fri Oct 13, 2017 2:45 am

I've got another 1700 sample (added to graph) - a warranty replacement for the performance marginality problem. At 2.7 GHz it's basically still idling, and it follows The Stilt's measurements on up. It hardly burns power and is generally amazing. If AMD has a lot of dies like this for servers and laptops, they've got it made.

Who is online

Users browsing this forum: blitzy, Kennethtoimb, Yan and 11 guests