Monday, March 13th 2017

AMD Says Ryzen 1700X, 1800X Have a Temperature Reporting "Offset"

AMD is now saying reports of poor thermal performance from the flagship Ryzen products can be attributed to a simple thing: Temperature Offsets. Apparently, to keep a "consistent fan policy," AMD has placed a 20C offset on the Ryzen 1700X and 1800X products, making them report temperature a good 20C above what the sensor reads. This interesting design choice may most assuredly be confusing to end users, but AMD is confident software will soon automatically adjust for this offset and report the true temperature when required.

In the same blog post detailing the changes on the 1700X and 1800X, AMD claims that temperature reporting "may be offset on certain CPU models so that all models on the AM4 Platform have the same maximum tCTL value." This could mean other future models would utilize a similar setup, so remember that moving forward with AMD's Zen-based lineup.
Source: AMD
Add your own comment

89 Comments on AMD Says Ryzen 1700X, 1800X Have a Temperature Reporting "Offset"

#26
cadaveca
My name is Dave
OneCoolHow much are you having to increase it from stock to get that?!
not that much. below 1.375V. Since temps are so high, I can't add more.
Posted on Reply
#27
mcraygsx
cadavecaWait, so my OC'd 67C load temps @ a bit over 200W are actually 47C? I doubt it. There is more to this story than meets the eye, because based on my load temps and the cooler I am using, temps should be a bit higher, perhaps.
Hey Dave, I just tested the temps for my 1800X w/BIOS 0902 for Crosshair VI Hero. Idle temps are far better then they were on the old BIOS. I used to sit idle at no less then 44c. But after this update IDLE temps seem to remain calm at around 36c for. I can confirmed this with BIOS and HWMonitor.

Posted on Reply
#28
cadaveca
My name is Dave
My idle temps are in the 20's. Were the same on even the original pre-launch BIOS.
Posted on Reply
#29
OneCool
Something tells me ambient temperature is making a lot of difference here.
Posted on Reply
#30
R-T-B
Ferrum MasterI wonder why multisocket CPU boards didn't show this before. They should suffer from the same problems actually.
Probably because they use NUMA or whatever it's called to mark themselves as separate.
Posted on Reply
#31
cdawall
where the hell are my stars
I was seeing 60C load temps under water at 4.1. I really doubt with the radiator setup that it was 20C cooler than that...
Posted on Reply
#32
R-T-B
cdawallI was seeing 60C load temps under water at 4.1. I really doubt with the radiator setup that it was 20C cooler than that...
And even if it was, I think we can all agree: Setting the temp at an offset to control the "fan profile" of a cooler many may not even be using is quite foolish.
Posted on Reply
#33
TheGuruStud
Dippyskoodlez1.) I don't see why frequency is relevant. Frequency is an aspect of design, not performance.

2.) Why can't AMD have 'silicon as good as intel"? They walloped the shit out of them with the Athlon Xp and the base Athlon 64 run(By making really good design decisions). This is an excuse for a lackluster product that is simply compensating for poor design choices, rather than holding their feet to the fire like people hold Intel to.



Multi socket boards do 'show this issue', but NUMA is important because the memory controllers are physically separated from CPUs(N.on U.niform M.emory A.ccess). Numa issues do not apply to Ryzen because they are all bound to the same pair of memory controllers. (Besides, Windows has been NUMA aware for a long, long time at this point)

ccnuma is a potential issue, but re: scheduling, AMD has cleared the air entirely of all scheduler issues. There is nothing to fix.

The real issue is just the reality of using a low speed CCX between quads for AMD.

community.amd.com/community/gaming/blog/2017/03/13/amd-ryzen-community-update?sf62109582=1
LOL, nothing to fix. Sure.

Hey, windows, how about you load two heavy threads on the same core, b/c you don't know the difference between logical and physical. It's totally going to run the same speed as putting each thread on real cores. Dippydingaling told me so!

Scale that up to more threads, plus, spread across the CCXs when the relevant threads could be put on one CCX.

AMD has MORONS speaking.

BTW, didn't anyone see benchmarks with a CCX shut off and SMT disabled? Magically, the bug is gone! AMAZING! lol
Posted on Reply
#34
Totally
OneMoar20c what consistent fan policy what .... in the actual frigg
this smells of amds marketing spin

why TF would you adjust the sensor +20 fracking C to tune the fan profile
what are fan curves too hard for AMD engineers to understand go figure
there is no fracking way there is that much of a difference chip2chip
My assumption, there is a weak link on the chip that they can't read temps directly from that gets really hot. So while the chip is reporting x degrees that other part/area is x + <= 20 degrees.
Posted on Reply
#35
YukikazeQ
cadavecaMy idle temps are in the 20's. Were the same on even the original pre-launch BIOS.
If it's anything like FX the idle temps are completely inaccurate only temps under load matter because it's using a far more accurate way of calculating the temps in the silicon rather than a typical sensor.
Posted on Reply
#36
YukikazeQ
cadavecaWait, so my OC'd 67C load temps @ a bit over 200W are actually 47C? I doubt it. There is more to this story than meets the eye, because based on my load temps and the cooler I am using, temps should be a bit higher, perhaps.
It's possible the software you're using to read it already has an offset in place. Also because AMD doesn't use toothpaste as TIM (the silicon is properly soldered to the IHS) it would run a lot cooler than you may be used to seeing on poor quality Intel products
Posted on Reply
#37
nem..
cadavecaIt doesn't when you see the power consumption. my 1700X pulls well over 200W @ 3.8 GHz. It seems that the chips are running into a frequency wall because of power consumption from what I see.
4C/8 threads 7700k - 80w STOCK



8c/16 threads 1800x - 56w


8c/16 threads 6900k
Posted on Reply
#38
DarkHill
YukikazeQIt's possible the software you're using to read it already has an offset in place. Also because AMD doesn't use toothpaste as TIM (the silicon is properly soldered to the IHS) it would run a lot cooler than you may be used to seeing on poor quality Intel products
You really need to step down from your red horse.

Intel uses TIM instead of solder on their mainstream products only. And they do this because they can. the 7700k still has OC headroom with the TIM. Soldering it would reduce temps sure, but not provide much more OC headroom. Delidding has shown this. You might gain 1-200Mhz more, but the real gain is in the load-temps.

However the cpu is rated at a Tjunc of 100c so there is no need for them to use a more expensive method of cooling it since the TIM solution keeps it WELL within the specs as it is.

AMD use solder on ryzen because it needs to be, just like the 2011 chips from Intel is soldered.
Posted on Reply
#39
cadaveca
My name is Dave
nem..8c/16 threads 1800x - 56w
That's great that they got a nice CPU, but I got a 1700X, and it pulls 100W via 8-pin at stock. So I fail to see what your point is... :p I can easily provide pictures, video, whatever showing my CPU's power consumption.

Funny how TH gets a 95W CPU that only pulls 56W... maybe they measured the 65W 1700 and got their numbers mixed up. o_O
Posted on Reply
#40
londiste
how about linking to the reviews so we wouldn't have to google to find what the conditions were.
short version: all the graphs are from different conditions and are not comparable.

7700k
gaming - watch dogs 2, custom high cpu load test sequence.
max - intel ptu

1800x
gaming - metro ll (rather lean on cpu by todays standards)
max - luxrender/prime95 (prime afaik is still not working properly on ryzen)

6900k
you didn't post the right graph for gaming. that would be the one with 64.7 avg/73.3 max
max - fpu torture (i'd assume aida64). notably, 7700k had far lower consumption with fpu torture than ptu.
nem..4C/8 threads 7700k - 80w STOCK
8c/16 threads 1800x - 56w
8c/16 threads 6900k
Posted on Reply
#41
GreiverBlade
cadavecaMy idle temps are in the 20's. Were the same on even the original pre-launch BIOS.
maybe it's me but i fail to see the issue...

you said 67°c typical load OC @200W and ~middle 20°C idle and it's ... bad? (or did i just cross read it ? )
my 6600K at 4.4 idle at middle 20C° and typical load between 55 and 68°C depending on the game/software ...
actually that a 8 core 16 thread (albeit at a lower clock) manage to get the same temp as a quad ... 95W (oh wait OC too soo probably more than 95W )

though you mentioned not being able to add more V than 1.375 due to the temperatures ... you mean 67°? (because 67° load for me is kinda low )

price/performance Ratio... Ryzen take the Crown .... a 7700K is less and less desirable for me ... even a 6900K, a simple R7 1700 will suffice ... now i just wait to see a TPU ASRock Taichi review ...

also gaming wise ... i fail to see the fail ... even at 1080p ... all comparison i saw put the 1700 on par with a 7700K (but ... 3.8ghz vs 5.0ghz) with a +/-10fps margin depending the game for the same price ... i go with AMD next (that will not be soon nonetheless) and i know it's 4C/8T vs 8C/16T.

tho i am glad you called Ryzen R7 "mainstream" ... that comfort me :D
Posted on Reply
#42
notb
YukikazeQIt's possible the software you're using to read it already has an offset in place. Also because AMD doesn't use toothpaste as TIM (the silicon is properly soldered to the IHS) it would run a lot cooler than you may be used to seeing on poor quality Intel products
The "poor quality product" doesn't need soldering to stay way under it's thermal limits.
Don't you think criticising Intel for not using costly solutions to problems that don't exist is a bit unfair? :)

Sure, Intel could solder the IHS for another $5 or something, but why do it?
Before we know you'll be expecting IHS made of diamond, because using copper is to cheap, obvious and lazy.
Posted on Reply
#43
londiste
notbThe "poor quality product" doesn't need soldering to stay way under it's thermal limits.
Don't you think criticising Intel for not using costly solutions to problems that don't exist is a bit unfair? :)

Sure, Intel could solder the IHS for another $5 or something, but why do it?
Before we know you'll be expecting IHS made of diamond, because using copper is to cheap, obvious and lazy.
have you owned a skylake/kabylake intel cpu? haswell was also affected to a lesser degree.

i seem to have a rather unfortunate (but not the worst) 6700k specimen. the initial motherboard i got (gigabyte z170n-wifi which can burn in hell) had a hard throttle point at 60c. any kind of load whatsoever sent the cpu over that, with a ekwb predator 240 attached to it, this should really not have been the case.

at bone stock it will even now go 70c+ easily with heavy load. that is with watercooling that should be way more than needed to get rid of the heat. cpu block is good enough, it worked fine-ish with haswell (<60c) and it was simply awesome with sandy bridge (40c-ish). among these, skylake uses the least power...

there is no gradual warming up of the processor. even considering the small size of the cpu die/ihs, the jumps in temperature are way too fast. idles at 35-40, something causes load and cpu gets +20c or more just like that. again, this was not how it worked with sandy and to lesser degree, haswell.

tl;dr - intel's tim can go to hell. soldering ftw.
Posted on Reply
#44
SPLWF
Does this offset occur in the FX lineup as well? I'm curious because my other gaming PC, my 8320 reports about 50-60c in the bios. When gaming it hits 72c (using HW Monitor). I was under the impression that 62-65c was tjMax. My 8320 is using an Cooler Master 120mm AIO. Thanks in advance.
Posted on Reply
#45
YukikazeQ
notbThe "poor quality product" doesn't need soldering to stay way under it's thermal limits.
Don't you think criticising Intel for not using costly solutions to problems that don't exist is a bit unfair? :)
I don't think it's unfair because even their highest end 'gaming/mainstream' overclockable(that you pay extra for to be able to do) 7700K is using the same toothpaste and runs close to 100C when overclocked with the stock toothpaste and 25C+ lower by deliding and replacing the toothpaste with TIM that isn't quite as good as being directly soldered on.
Posted on Reply
#46
YukikazeQ
SPLWFDoes this offset occur in the FX lineup as well? I'm curious because my other gaming PC, my 8320 reports about 50-60c in the bios. When gaming it hits 72c (using HW Monitor). I was under the impression that 62-65c was tjMax. My 8320 is using an Cooler Master 120mm AIO. Thanks in advance.
FX chips don't have an offset as far as I know but use the same method of calculating the temp directly in silicon as Ryzen. The only software I've seen AMD reps recommend that reads the temps for FX correctly was overdrive/core temp. The t.j.max for most FX chips is 80C.
Posted on Reply
#47
alucasa
cadavecaIt doesn't when you see the power consumption. my 1700X pulls well over 200W @ 3.8 GHz. It seems that the chips are running into a frequency wall because of power consumption from what I see.
U serious on power consumption?
Posted on Reply
#48
cadaveca
My name is Dave
alucasaU serious on power consumption?
"Why wouldn't I be" perhaps is the more prudent question.
Posted on Reply
#49
cdawall
where the hell are my stars
I saw similar, but not as high from the 1800X's I tested with.
Posted on Reply
#50
SPLWF
YukikazeQFX chips don't have an offset as far as I know but use the same method of calculating the temp directly in silicon as Ryzen. The only software I've seen AMD reps recommend that reads the temps for FX correctly was overdrive/core temp. The t.j.max for most FX chips is 80C.
Thanks, I will install core temp/overdrive and compare temps. Thanks again :)
Posted on Reply
Add your own comment
Dec 22nd, 2024 14:28 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts