Is it normal for the voltage to oscillate in between around .05 V during normal operation? Why would it do that?
Slightly unrelated, but what are SYSTIN, CPUTIN and AUXTIN on the temperature reporting tool?
Is it normal for the voltage to oscillate in between around .05 V during normal operation? Why would it do that?
Slightly unrelated, but what are SYSTIN, CPUTIN and AUXTIN on the temperature reporting tool?
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
The only good reason for a voltage to be jumping around would be if your processor has some sort of power saving operations.
What temperature tool are you using? I'd like to know what the values on those are... it'd help me guess what each represents. Generally, though... anything SYS would be your northbridge, CPU would obviously be your processor, and AUX would be referring to temperature readings on your PCIe sockets. You may have AUX0, AUX1, AUX2, etc labels. At least this is the case on any temperature tool that used these labels.
It's worth noting that if CPUTIN is a temperature, then it would be a reading closer to the temperature of the IHS and not the reading of the cores themselves. Those should be listed seperately as Core0, Core1, etc...
Last edited by SlyMaelstrom; 06-10-2008 at 08:16 PM.
Sent from my iPadŽ
I'm using the CPUID tools (http://www.cpuid.com/hwmonitor.php).
Below is a screenshot. The actual variance is low there. But it usually ends up in .05 V. Haven't seen it any higher than that.
I'm not getting discrete AUX readings, but then and again I only have one slot occupied... The high AUX temperature then makes sense, considering that's where the graphics card is sitting... Good to know these things
Attachment 8203
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
It's strange to see your core temps lower than that CPUTIN row. The variance between the AUXTIN and the GPU core make me doubt they're related, as well. However, the data I'm questioning is the one on the GPU Core row as those are extremely high if you're just sitting there on your desktop. The AUXTIN looks like a more accurate number even for the passively cooled model.
I mean... when I download the tool, everything looks about right. To the point that I would say the CPUTIN is in fact the temp around the IHS... but I still find it abnormal that your cores have a lower temp than the IHS by a fairly significant amount. I should also mention that my AUXTIN is reading at 127 °C, which is clearly out of whack, so I have no idea what it's reading.
Heh, anyway... I'd like to post a pic of mine but the temps are embarrassing with the heatwave we've been having in New York and the fact that my air conditioner is broke. I'm thinking I will put the processor back at stock clocks until it cools down. Anyway, the variance in the voltage isn't that bad. I couldn't say for sure but I think .05v is acceptable. I'd watch it if it goes more than that, though.
Last edited by SlyMaelstrom; 06-10-2008 at 10:43 PM.
Sent from my iPadŽ
I would say that 50mV (0.05V) variation in the voltage in is perfectly normal. The 3.3V power comes directly from the PSU, and it's known to not be entirely stable. The one you don't want to vary too much is the CPU VCore - at least not unless you have set up the system to do frequency/voltage changing based on load.
There is no telling what the temperature sensors are actually connected to - it's up to the individual motherboard vendor to use those sensors as they see fit - and the one saying CPU may well be a sensor somewhere near the CPU - it is obviously not the CPU core itself, because that's reported separately elsewhere. An AUXTIN of 127'C would indicate that it's a "not connected" input on the temperature sensor chip.
--
Mats
Compilers can produce warnings - make the compiler programmers happy: Use them!
Please don't PM me for help - and no, I don't do help over instant messengers.
Ah, ok. Excellent.
The CPU VCore is exactly the one I'm looking and reporting a 0.5 V variance (despite being lower on that particular screenshot).
Meanwhile, the GPU is indeed high. This relates to a recent thread here on the Geforce 7300. Basically the card is fanless and NVidia assured me (along with you folks) that's an acceptable temperature, being only 120c a worrisome number as Elysia mentioned.
I'll try and check other reporting tools and compare results, nonetheless.
Over her in Portugal, spring is coming late, it's coming fast and it's coming furious. Temperatures are already on the 29c (84f) mark, having raised from 16c (61f) just last week. Until the week ends we are expected a consistent 35c (95f). An this is not even summer yet. This is going to be a bad year for my heating concerns.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
Is it varying more than 50mV from the median value, or is it 50mV from peak to peak. The latter is well within the margins, the former just about OK [at least in AMD's processors, and I would expect that Intel specs are similar if not the same].
--
Mats
Compilers can produce warnings - make the compiler programmers happy: Use them!
Please don't PM me for help - and no, I don't do help over instant messengers.
Originally Posted by brewbuck:
Reimplementing a large system in another language to get a 25% performance boost is nonsense. It would be cheaper to just get a computer which is 25% faster.
With that 7200 GS it will be...
When we didn't have air conditioning I tried to avoid turning on my PC when my room was above 45C (yes it gets that hot).
Idle? No, no that's not even close to true. I think maybe you meant the physically acceptable temps for the cards. However, even at load you shouldn't exceed 80 °C on any graphics card. You really want to try to keep it under 70°C even. I don't know where you got your information, but please let's not try to spread it around and confuse people.
Here are some reviews of some nVidia cards with temperature benchmarks:
http://www.overclockersclub.com/revi...atal1ty/10.htm
http://www.nvnews.net/reviews/evga_g...c/page_9.shtml
http://www.legitreviews.com/article/384/10/
There is a big difference between what a GPU can withstand and what is safe for it. I think all of these GPUs have a physical threshold of 130 °C, but you really don't want to go anywhere near that. If you keep letting it hover around where it is, you're surely going to be replacing it much sooner than you'd ever imagine. Really, the only reason you should be exceeding that 80 °C temperature on your normal card is if it's being overclocked. Then, I would say keep it below that 100 °C point.
Last edited by SlyMaelstrom; 06-11-2008 at 10:01 AM.
Sent from my iPadŽ
I didn't quote Mario's statement about nVidia claiming that to be an acceptable temperature for the fanless card (even though, I'm quite sure they didn't intend that to mean an idle temperature). I quoted your statement about being equal to above 100 °C on idle as normal for newer nVidia cards. That is an absurd statement and if you really wanted to email nVidia, they would agree with that.
The fact is, all nVidia cards will start to throttle at 110 °C, so if it's going to hover around 100 °C at anything but peak load, then when you do reach peak load, you're only going to get reduced performance. Anyway, the temperature that the original picture Mario posted is showing a temperature about 15 °C above what a fanned 7 series card would show. For a fanless, I suppose this is acceptable, but those cards are never designed to reach a peak load and I would be careful if you intend to do anything that would put that much stress on it.
Sent from my iPadŽ