Ethereum Community Forum
Home› Mining
Vega Finally Unleashed! Fine-tuned performance and power consumption in a single chassis.

PSA: RX480 and unpowered risers (exceeds 75w ATX spec)

LogicaluserLogicaluser Member Posts: 214 ✭✭
edited June 2016 in Mining
Initial reports from reviews indicate that RX 480 draw considerable power from the PCI-E slot
(I'm talking about +75w, with short-term peaks of +100w)

Until we have some solid measurements under mining conditions, I would caution everyone to avoid using unpowered ribbon risers, since this is anywhere from 2-5x the wattage typical mining cards pull from PCI-E slot, more than double that drawn by a MSI 290X Lightning.
Hell, I would even caution against direct-mount triple card setups at this point.

RX480 82w average from PCI-E slot
http://www.tomshardware.com/reviews/amd-radeon-rx-480-polaris-10,4616-9.html

MSI R9 290x Lightning 32w average from PCI-E slot
http://www.tomshardware.com/reviews/r9-290x-lightning-performance-review,3782-5.html

Sapphire R 380X Nitro 48w average from PCI-E slot
http://www.tomshardware.com/reviews/amd-radeon-r9-380x-nitro,4361-6.html
· Share on TwitterShare on Facebook
«1

Comments

  • Topr3Topr3 Member Posts: 23 ✭
    Already ordered 6x rx 480, have powered risers... hope everything stays alive :/
    · Share on TwitterShare on Facebook
  • IkeagoIkeago Member Posts: 130 ✭
    I'd be careful how you run the power to your risers.

    If your running off a consumer PSU I would not run more than 2 cards on 1 6 pin cable.
    · Share on TwitterShare on Facebook
  • Topr3Topr3 Member Posts: 23 ✭
    Got 6x6pins evga 1000w t2, I think I'll be fine, right?
    · Share on TwitterShare on Facebook
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    @Topr3 1000w isn't going to be enough for 6x 480s at stock clocks and voltage.
    · Share on TwitterShare on Facebook
  • Topr3Topr3 Member Posts: 23 ✭
    dlehenky said:

    @Topr3 1000w isn't going to be enough for 6x 480s at stock clocks and voltage.

    Will play with it if needed, and it is a titanium rated psu so won't be so far from doing it
    · Share on TwitterShare on Facebook
  • IkeagoIkeago Member Posts: 130 ✭
    I doubt you'll be able to get all 6 working but you won't melt any wires at 1 wire per riser.
    · Share on TwitterShare on Facebook
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    @Topr3 With stock settings, you'll be pushing it with a 1200w, titanium or not. I run 6 Nanos, so ... I have EVGA G2 GOLD 1300w and the stock Nanos just fit. With undervolting (thank you @Heliox!!) they're down around 1050w with overclocking. These numbers are all at the wall, so they include all the other system components, too.
    · Share on TwitterShare on Facebook
  • HelioxHeliox Member, Moderator Posts: 634 mod
    :) Glad you're happy @dlehenky, it's awesome to get some appreciation here and there! That's all i want/need!

    Next step: 480's ;). Tomorrow is the big day! :)

    On topic: Powered risers will sort out that issue for sure!
    That's the whole idea behind those powered risers.. don't overdue those pci-lanes :)

    Will post some figures "hopefully" this weekend!

    Greetings!
    · Share on TwitterShare on Facebook
  • Marvell9Marvell9 Member Posts: 593 ✭✭✭
    Ikeago said:

    I doubt you'll be able to get all 6 working but you won't melt any wires at 1 wire per riser.

    I've run six nanos dual mining at 136 watts on a single 1000watt platinum evga, with a 100 mv undervolt you should be able to do six 480s on 1000 watt i think
    · Share on TwitterShare on Facebook
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    edited June 2016
    @Heliox even with powered risers it's a lot more than most people typically push through molex, I can imagine plenty of burnt components as people try in-place replacements for Tahiti cards not realizing it's +3x the draw via the risers molex/sata connections.

    @Marvell9 I highly doubt Nanos are pulling 70-80w per card avg via the PCI-E slot like these RX480 do at stock.

    6 cards may very well be 450w total being drawn through risers & their Molex (or worse, SATA adapters), quite a different situation from the cards we're used to in past generations that sip 10-30w through the PCI-E slot and get the bulk of their power via their 6/8pins
    · Share on TwitterShare on Facebook
  • weirdgodweirdgod Member Posts: 3 ✭
    This is nice news for all the buyers of 4GB rx480 cards... Looks like we'll be able to get "free upgrade" to 8GB version once someone dumps a bios from 8gb :)

    http://www.legitreviews.com/amd-radeon-rx-480-4gb-video-cards-8gb-memory_183548

    And since same memory chips are used, i think overclocking the mem towards 9Gh is plausible.

    I have few Powercolor's 4GB RX480 coming in in a few days... can someone dump a bios from 8GB version (preferably powercolor, but at this reference design point, i am sure any would work)? I'll try flashing one to 8GB :D
    · Share on TwitterShare on Facebook
  • boysieboysie Member Posts: 591 ✭✭✭
    weirdgod said:

    This is nice news for all the buyers of 4GB rx480 cards... Looks like we'll be able to get "free upgrade" to 8GB version once someone dumps a bios from 8gb :)

    http://www.legitreviews.com/amd-radeon-rx-480-4gb-video-cards-8gb-memory_183548

    And since same memory chips are used, i think overclocking the mem towards 9Gh is plausible.

    I have few Powercolor's 4GB RX480 coming in in a few days... can someone dump a bios from 8GB version (preferably powercolor, but at this reference design point, i am sure any would work)? I'll try flashing one to 8GB :D

    haha that's awesome...esp as I got 4 of the 4GB cards...surprise...I actually bought 8GB cards priced and bios flashed to limit it to 4GB lols...

    Please post how and when we can flash these to 8GB :) although as long as the speed is increasable to the that same as the 8GB why flash at all? or is the bios also limiting the speed?

    Boysie
    · Share on TwitterShare on Facebook
  • Wolf0Wolf0 Member Posts: 329 ✭✭✭
    Actually, PCI-E 3.0, which is what it's for, specifies 150W from the slot.
    · Share on TwitterShare on Facebook
  • workwork Member Posts: 2,084 ✭✭✭✭
    edited June 2016
    @Wolf0 pcisig.org says "75W from x16 PCIe connector." 150W still requires a 6pin connector.

    @weirdgod now I'm super sad I didn't manage to snag any 4GB cards. Appears they haven't been imported to a lot of places yet - but, at least I kinda get why now.
    · Share on TwitterShare on Facebook
  • Wolf0Wolf0 Member Posts: 329 ✭✭✭
    work said:

    @Wolf0 pcisig.org says "75W from x16 PCIe connector." 150W still requires a 6pin connector.

    @weirdgod now I'm super sad I didn't manage to snag any 4GB cards. Appears they haven't been imported to a lot of places yet - but, at least I kinda get why now.

    I remember reading the new spec... but oh, well. Maybe you're right.
    · Share on TwitterShare on Facebook
  • bbcoinbbcoin Member Posts: 377 ✭✭✭
    edited June 2016
    boysie said:

    weirdgod said:

    This is nice news for all the buyers of 4GB rx480 cards... Looks like we'll be able to get "free upgrade" to 8GB version once someone dumps a bios from 8gb :)

    http://www.legitreviews.com/amd-radeon-rx-480-4gb-video-cards-8gb-memory_183548

    And since same memory chips are used, i think overclocking the mem towards 9Gh is plausible.

    I have few Powercolor's 4GB RX480 coming in in a few days... can someone dump a bios from 8GB version (preferably powercolor, but at this reference design point, i am sure any would work)? I'll try flashing one to 8GB :D

    haha that's awesome...esp as I got 4 of the 4GB cards...surprise...I actually bought 8GB cards priced and bios flashed to limit it to 4GB lols...

    Please post how and when we can flash these to 8GB :) although as long as the speed is increasable to the that same as the 8GB why flash at all? or is the bios also limiting the speed?

    Boysie
    Did it occur to you it might have been a Press review PCB to make guys like you jump on the 4GB version ;) ?
    Are you willing to open the card and actually check wheter you got the 8gb of memory?
    One way or another BIOS flash would cancel your warranty.
    · Share on TwitterShare on Facebook
  • peepeedogpeepeedog Member Posts: 32 ✭
    Well, I may be the first person to fry my H81 Pro BTC with unpowered risers.

    Got my RX480 today and just plugged in - installed the latest drivers and ran Claymore's dual miner.

    Got a nice 24 hashrate at stock speeds, but after 3-4 mins, machine rebooted on it's own, and failed to post.

    Removed the RX480, the machine still won't post :\

    Wheeeeee - please use powered risers
    · Share on TwitterShare on Facebook
  • Marvell9Marvell9 Member Posts: 593 ✭✭✭
    why in gods name do you use unpowered risers?
    · Share on TwitterShare on Facebook
  • peepeedogpeepeedog Member Posts: 32 ✭
    Marvell9 said:

    why in gods name do you use unpowered risers?

    Because a properly spec-ed card won't cause issues?

    I've been mining since January using unpowered risers with 3 x 280X, 1 x 7970 and 1 x 7990 with no issues whatsoever. But yeah with the new 480s i'm switching all the cards to powered.

    · Share on TwitterShare on Facebook
  • Marvell9Marvell9 Member Posts: 593 ✭✭✭
    one time i had an upowerd riser cheapie from china catch fire luckily i was there when t happend
    unpwered risers are not worth the risk
    · Share on TwitterShare on Facebook
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    Damn, I knew this was not going to go well.

    @peepeedog Tahiti-based cards draw like 60-75% less through the PCI-E slot than a RX 480 does.

    Swap 6x 7970 with 6x 480 and you literally quadruple the power draw via risers, even though you cut overall power usage by +30%

    I mean shit... even using powered USB style risers, I'd imagine many have rigs with 3 risers off each modular molex strand, or which are using SATA power with those thin little adapters (rated for 20-25w iirc)
    · Share on TwitterShare on Facebook
  • peepeedogpeepeedog Member Posts: 32 ✭
    Logicaluser said:

    Damn, I knew this was not going to go well.

    @peepeedog Tahiti-based cards draw like 60-75% less through the PCI-E slot than a RX 480 does.

    Swap 6x 7970 with 6x 480 and you literally quadruple the power draw via risers, even though you cut overall power usage by +30%

    I mean shit... even using powered USB style risers, I'd imagine many have rigs with 3 risers off each modular molex strand, or which are using SATA power with those thin little adapters (rated for 20-25w iirc)

    Yeah agree, i'm just shocked that it died after using ONE RX 480! Wonder how it's going to survive normal gaming on lesser motherboards, i hear the witcher 3 does quite a bit of peak on the PCIE slot during gameplay.

    Anyhow, learnt my lesson!

    Waiting for the replacement board and powered risers to come in, and perhaps we should all wait for AIB cards with beefier 8pin connections.

    Cheers
    Shaun
    · Share on TwitterShare on Facebook
  • th00berth00ber Member Posts: 213 ✭✭
    Personnaly, I plan to swith my 7970 for RX 480, but I'll wait for custom card with a 8-pin connectors or dual 6-pins.
    150W for the card with 75W from the slot, even with powered riser, it sound a little bit risky
    · Share on TwitterShare on Facebook
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    @peepeedog I'm curious. Was this a X16 or an X1 ribbon riser that you were using?
    Any sign of damage to the riser or board?
    · Share on TwitterShare on Facebook
  • LogicaluserLogicaluser Member Posts: 214 ✭✭
    .
    · Share on TwitterShare on Facebook
  • peepeedogpeepeedog Member Posts: 32 ✭
    Logicaluser said:

    @peepeedog I'm curious. Was this a X16 or an X1 ribbon riser that you were using?
    Any sign of damage to the riser or board?


    This was an X1 riser, perhaps the X16 might have survived. No physical signs that i can see, either on the board or the 1X connector. Most reports have it that the PCIE slot dies, but the other slots are working. In my case I can't even post - no bios, nothing at all. It goes into an infinite loop of restarting - fans spin up, nothing on screen, then shuts down after awhile - then repeat.

    I've tried removing and replacing the battery on the mobo thinking it might clear the CMOS or something, but no luck there either.

    It's an old H81 BTC Pro, so i'm not too fussed (plus my folly of running it off an unpowered PCIE slot). I've ordered the H97 anniversary, will test with that and report back.

    At least AMD did come out to say something to the effect of acknowledging the issue.

    Cheers
    Shaun
    · Share on TwitterShare on Facebook
  • HelioxHeliox Member, Moderator Posts: 634 mod
    Here

    Firstly: This fix is by no means an official one and you will be using it at your own risk. If you for any reason manage to fry your card while using it, you are not entitled to make any claims for warranty exchange (RMA).

    As soon as the bios flashing software supporting Ellesmere becomes available, I'll be releasing a modified bios which makes the necessary changes automatically.

    The procedure is rather simple, and can be done on any software which supports I2C control on Radeon graphics. Personally I recommend MSI Afterburner for the job.

    - Download and install the newest MSI Afterburner
    - Go to the installation folder of MSI Afterburner (where the MSIAfterburner.exe is located) and open a new command window ("Left Shift + Right Mouse Click" when no file is selected)
    - Additional step: Verify that the I2C interface to the VRM controller is working by typing following command: "MSIAfterburner /ri06,08,0D" (without the quotes). If the response is ": 20" or ":44" you're good to go.
    - The two commands you need to input are following: "MSIAfterburner /wi06,08,1E,BB" and "MSIAfterburner /wi06,08,1F,B0".

    Once you have given these commands the power draw from the PCI-E slot has reduced and moved towards the PCI-E power connector. The recommended programming value is "B" (/wi06,08,1E,BB and /wi06,08,1F,B0), however you can go set it to anywhere between "B" and "F". Higher the value, more of the power draw will be directed to the PCI-E power connector.

    Do note that this fix must be re-applied after every reboot.

    EDIT: Added the I2C response for IR3567B controller (0x44).
    IR3567A = 0x20.

    The fix applies for both!



    Shout out to the stilt!
    · Share on TwitterShare on Facebook
  • bbcoinbbcoin Member Posts: 377 ✭✭✭
    edited July 2016
    I just got my 4x 480, unpowered ribbon risers caused crashes. As soon as I connceted them to molex everything runs fine... So yeah, 480 defo needs powered risers.
    · Share on TwitterShare on Facebook
  • dlehenkydlehenky Member Posts: 2,249 ✭✭✭✭
    What I don't get is that the rx 480, a 14 nm process chip, is drawing as much power as the Nano, which is 28 nm process chip. It should have been a slam-dunk for AMD to get *magical* power reduction numbers for the rx 480, but apparently not. I can only assume, at this point, that the rx 480 represent AMD's first-pass effort at moving to the 14 nm chip process, and as such, the chip design is anything but optimized for the new process. In other words, it appears to me that the rx 480 is a "let's get something out the door quick" offering, and doesn't show anything near what can be accomplished on a robust/mature 14 nm process chip design. I'd call it a rev 0.9 type effort, if that.
    · Share on TwitterShare on Facebook
  • bbcoinbbcoin Member Posts: 377 ✭✭✭
    edited July 2016
    dlehenky said:

    What I don't get is that the rx 480, a 14 nm process chip, is drawing as much power as the Nano, which is 28 nm process chip. It should have been a slam-dunk for AMD to get *magical* power reduction numbers for the rx 480, but apparently not. I can only assume, at this point, that the rx 480 represent AMD's first-pass effort at moving to the 14 nm chip process, and as such, the chip design is anything but optimized for the new process. In other words, it appears to me that the rx 480 is a "let's get something out the door quick" offering, and doesn't show anything near what can be accomplished on a robust/mature 14 nm process chip design. I'd call it a rev 0.9 type effort, if that.

    Sure but the price of the Nano proves it's very expensive to make 'eco' stuff. Or at least eco = charge premium at the moment.
    Same with solar power and electric cars. That stuff is still expensive and will take time to make sense for everyday folks.

    I'm actually glad that MORE POWER! for the sake of power race is over. No more 8ghz cpu's and 1000W power supplies at home.No more R9 295x2 burning cables and killing $20 power supplies.
    I mean when I saw someone using 1kw supply for the first time I remember saying to myself 'why?'

    Now we have an actual improvement where components perform better yet save almost 50% on electricity.

    PS. Corsair RM850 gold is the most powerful PSU I have ever used and it looks like it will be the most powerful ever :)
    · Share on TwitterShare on Facebook
Sign In or Register to comment.