• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

The 480: power consumption, PCI-E powerdraw

Page 15 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Silverforce11 said:
The first proven casualty of RX 480s.. 3 of them actually.

While mining ETH on a el-cheapo MB, on non-powered (!!) risers.

https://bitcointalk.org/index.php?topic=1433925.msg15438155#msg15438155

He made the noob mistake of mining multi GPU without powered-risers and paid the price. hehe

To be fair, this could happen for other multi-GPU mining setups as well and it did in the past. This is why miners pay a few $ extra for molex-powered risers to supply additional power to the GPU, rather than going through the MB.
Click to expand...

oh nice!!! With this report , No miner can abuse RX 480 just like R9 290X 😀
 
96Firebird said:
Looks like something got lost in the translation...

http://www.tomshardware.com/reviews/nvidia-geforce-gtx-1080-pascal,4572-10.html
Click to expand...

I need no translator to read french, THG s french version stipulate that this connector is certified for 125W, notice that they didnt say that it s the PCIe rating, so it s possible that it s this connector rating...

Le connecteur d'alimentation PCI-Express à 8 broches fournir 133 W, même s'il n'est certifié que pour 125 W.
Click to expand...
 
littleg said:
That's the peak power consumption. As I understand it the problem isn't with the peak power, it's a problem if the card is continuously drawing more power than the slot is rated for.

I believe the spec is 75W +/- 8% also which puts the 950 within the margins.
Click to expand...

Those type of spikes are much worse. Regardless, I don't care either way and frankly tired of debating it. Your computer won't die or blew up in either case.
 
torlen11cc said:
I don't understand that issue.
Many GPUs draw more than 75W, like 950 (no power connector):

power_peak.png


so... what is the problem with rx 480?
Click to expand...


This review wasn't hard to find.

https://www.techpowerup.com/reviews/ASUS/GTX_950/21.html


power_average.png


power_maximum.png



Do you understand why they also have charts for average and maximum while the one you're citing is peak?
 
Last edited:
Zstream said:
Those type of spikes are much worse. Regardless, I don't care either way and frankly tired of debating it. Your computer won't die or blew up in either case.
Click to expand...

Much worse than what? The RX 480 is drawing an average of 86W over the PCIe slot (and obviously peaking much higher), are you really saying that the 79W peak (with an average of 74W) of the Asus 960 as measured by TPU is worse than that?

Erenhardt said:
I bet it takes power from earth's gravity field to run those extra FPS
perf_oc.png
Click to expand...

You do understand that act of overclocking is the user specifically choosing to run their hardware out of spec. So it's hardly surprising when said hardware no longer conforms to specs.
 
Last edited:
86W wont kill anything. This doesn't matter unless you are like that guy putting 3 in his system for hours of mining
 
Azix said:
86W wont kill anything. This doesn't matter unless you are like that guy putting 3 in his system for hours of mining
Click to expand...

Again, MB designers like to use the ATX plug as a failure point in that aspect, its easier for them, as that can be repaired easily. 3 RX480 could destroy both the PSU and MB ATX plugs, may happen with 2 as well, specially if you also have other PCI-E devices.

But the thing is, you cant ensure a thing, its unlikely that anything gona happen with peak power, peak power may blow up MB components, but MBs can handle that i think, constant power is a problem because it can generate HEAT, it may destroy traces (highly unlikely), on a PCI-E 1.0 board it may heat up the pci-e slot pins, killing both the slot and the card, but even them i dont think its enoght to heat up the ATX plug, and thats by far the weakest spot. Again, unless you have several other pci-e devices, that uses a considerable amount of power.

The ATX plug does not even allow MB OEM to fully meet PCI-E 1.0 power if the MB has 3 pci-e slots, the one thing i can get from this is that we need a new ATX power standart that could at least supply about 600W of power to the MB. I think its time to drop ATX compatibility.
 
Last edited:
What's FP64 compute performance of RX480? Power consumption certainly seems high, especially if it's FP64 isn't high.
 
Last edited:
I'll be the first to post if my 2x RX 480 plugged directly into the motherboard cause any issues.

So far, for 100% mining load they haven't.
 
I am replicating "lazy user" settings as much as possible since there will always be someone around to monitor the rig in case something happens. They are throttling a bit at stock settings but that's to be expected with a dual card config in a regular ATX case.

The motherboard is a ASUS P8Z77-V LK with no auxiliary 12V power for the PCI-e slots. So overall, it should closely replicate the conditions alleged to have caused the failures, though admittedly I am "only" using two cards instead of three.
 
Shivansps said:
Check if the ATX plug gets hot on where the yellow wires are.
Click to expand...

I am familiar with the ATX pin layout and have tried feeling the +12V area of the connector for abnormal temperatures. Doesn't feel hot to me. Which is why I'm not worried about it.
 
Silverforce11 said:
The first proven casualty of RX 480s.. 3 of them actually.

While mining ETH on a el-cheapo MB, on non-powered (!!) risers.

https://bitcointalk.org/index.php?topic=1433925.msg15438155#msg15438155
Click to expand...
Wait a minute, that guy ran three cards off a mainboard that only has two PCIE x16 slots and two PCIE x1 slots? x1 slots are rated for 25W only (There's more than one reason for powered risers!). He went out of spec regardless of which GPUs he would use. A lot of GPUs would have killed his board eventually with that configuration and the triple 280x he used prior were just that low on the PCIE power load to make it work.
 
Shivansps said:
3 RX480 could destroy both the PSU and MB ATX plugs, may happen with 2 as well, specially if you also have other PCI-E devices.

.
Click to expand...

Complete non sense, the MB PCIe has a capability of 300W for a 4 slots MB, if a permanent drain of 240W could cause a problem just imagine the GTX960 in SLI (as often recommended here) that should drain 500W peaks for two cards.

Anyway your post is noticed so in a few weeks i can remind you how fuddy it was retrospectively...
 
Abwx said:
Complete non sense, the MB PCIe has a capability of 300W for a 4 slots MB, if a permanent drain of 240W could cause a problem just imagine the GTX960 in SLI (as often recommended here) that should drain 500W peaks for two cards.

Anyway your post is noticed so in a few weeks i can remind you how fuddy it was retrospectively...
Click to expand...

pcie slots that expect video cards can supply 75 watts.

x1, x4, x8 slots that do not expect video cards can do 25 watts. some boards allow more (the asrock btc board for example).

check the pcie specs.

please correct me if Im wrong.
 
Abwx said:
Complete non sense, the MB PCIe has a capability of 300W for a 4 slots MB, if a permanent drain of 240W could cause a problem just imagine the GTX960 in SLI (as often recommended here) that should drain 500W peaks for two cards.

Anyway your post is noticed so in a few weeks i can remind you how fuddy it was retrospectively...
Click to expand...

960's aren't burning up boards, 480's are. Fact. Making up your own PCIe standards won't change that.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top