Page 1 of 2 12 LastLast
Results 1 to 10 of 12

Thread: RTX3080 review by Igor's Lab

  1. #1
    Join Date
    Feb 2020
    Location
    Uzbekistan
    Posts
    361
    Thanks Thanks Given 
    68
    Thanks Thanks Received 
    82
    Thanked in
    74 Posts

    Default RTX3080 review by Igor's Lab

    https://www.igorslab.de/nvidia-gefor...n-fuer-turing/

    So it's essentially 400W GPU after all ...


    Also, instantaneous FPS is quite bad in some titles for some reason, early drivers i guess ?

  2. #2
    Join Date
    Aug 2010
    Posts
    2,742
    Thanks Thanks Given 
    131
    Thanks Thanks Received 
    6
    Thanked in
    6 Posts

    Default

    Not impressed tbh, higher performance yes, but also higher power consumption.

    And availability will suck for a while.

  3. #3
    Join Date
    Aug 2020
    Location
    ROK
    Posts
    19
    Thanks Thanks Given 
    0
    Thanks Thanks Received 
    3
    Thanked in
    2 Posts

    Default

    Quote Originally Posted by Juular View Post
    So it's essentially 400W GPU after all ...
    2020091612121493.jpg
    Attachment 4221
    Attachment 4222


    https://www.fcpowerup.com/nvidia-308...compatibility/

    However, Fantastic(edited)'s review says the RTX 3080 can consume 750 watts in a very short time.
    The total instantaneous power consumption of the system including the 9900k is 1000w.

    So he recommends 850w because he thinks these spikes are not good for the PSU.

    What do you think about this?
    Last edited by Barcode; 09-16-2020 at 03:45 PM.

  4. #4
    Join Date
    Aug 2018
    Location
    South Korea
    Posts
    518
    Thanks Thanks Given 
    203
    Thanks Thanks Received 
    65
    Thanked in
    41 Posts

    Default

    Quote Originally Posted by Barcode View Post


    https://www.fcpowerup.com/nvidia-308...compatibility/

    However, Fantasy's review says the RTX 3080 can consume 750 watts in a very short time.
    The total instantaneous power consumption of the system including the 9900k is 1000w.
    And he also claims that other reviewer's probe sampling rate is too low. For ex TH uses 1ms setting and he claims that's too long to capture very short spikes. I remember he mentioned igor's too.

  5. #5
    Join Date
    Oct 2006
    Location
    jonnyGURU forums, of course!
    Posts
    17,304
    Thanks Thanks Given 
    579
    Thanks Thanks Received 
    652
    Thanked in
    476 Posts

    Default

    Yes. Nvidia has published some specs showing 500us (MICRO second) spikes in power. Not sure if they gave that info to reviewers or just shared it with engineers. Given FCPOWERUP's "position" in the "industry" he's probably referencing the same "data" I have "seen".
    Last edited by Jon Gerow; 09-16-2020 at 01:43 PM.

  6. The Following User Says Thank You to Jon Gerow For This Useful Post:

    Philipus II (09-20-2020)

  7. #6
    Join Date
    Aug 2018
    Location
    South Korea
    Posts
    518
    Thanks Thanks Given 
    203
    Thanks Thanks Received 
    65
    Thanked in
    41 Posts

    Default

    Quote Originally Posted by Barcode View Post
    However, Fantasy's review says the RTX 3080 can consume 750 watts in a very short time.

    Quote Originally Posted by Jon Gerow View Post
    Yes. Nvidia has published some specs showing 500us (MICRO second) spikes in power. Not sure if they gave that info to reviewers or just shared it with engineers. Given Fantasy's "position" in the "industry" he's probably referencing the same "data" I have "seen".
    But his nickname is Fantastic, not Fantasy

    Looks like he measured data with his own equipment, he has chromas and high sampling rate probes/oscils.

  8. #7
    Join Date
    Oct 2006
    Location
    jonnyGURU forums, of course!
    Posts
    17,304
    Thanks Thanks Given 
    579
    Thanks Thanks Received 
    652
    Thanked in
    476 Posts

    Default

    Quote Originally Posted by Polaris20 View Post
    But his nickname is Fantastic, not Fantasy

    Looks like he measured data with his own equipment, he has chromas and high sampling rate probes/oscils.
    I get them confused. I changed it to "FCPOWERUP" in my reply.

    And it's not his own equipment. He has access to an OEM we're all quite familiar with. That's how he managed to leak those connector pictures so early.

    And having test equipment is one thing. Knowing what to test for is another. He wouldn't have tested < 1ms if he wasn't given direction to do so. We (Corsair) did the same test because of the same direction given by the same people. This is why Corsair says "at least an 850W" on their website instead of 750W like everyone else is suggesting. Thought the charge in the caps on the DC side alone should be able to sustain such short bursts of power (the 500us).

    750W is probably fine.. well.. most of them. We tested RM750 and SF750 and with a 70A 200us pulse repeated, both passed. When we increased that spike to 85A for 200us, however, the SF750 shut down. The RM750 eventually shut down when we increased that pulse to 106.25A.

  9. #8
    Join Date
    Aug 2018
    Location
    South Korea
    Posts
    518
    Thanks Thanks Given 
    203
    Thanks Thanks Received 
    65
    Thanked in
    41 Posts

    Default

    Quote Originally Posted by Jon Gerow View Post
    I get them confused. I changed it to "FCPOWERUP" in my reply.

    And it's not his own equipment. He has access to an OEM we're all quite familiar with. That's how he managed to leak those connector pictures so early.

    And having test equipment is one thing. Knowing what to test for is another. He wouldn't have tested < 1ms if he wasn't given direction to do so. We (Corsair) did the same test because of the same direction given by the same people. This is why Corsair says "at least an 850W" on their website instead of 750W like everyone else is suggesting. Thought the charge in the caps on the DC side alone should be able to sustain such short bursts of power (the 500us).

    750W is probably fine.. well.. most of them. We tested RM750 and SF750 and with a 70A 200us pulse repeated, both passed. When we increased that spike to 85A for 200us, however, the SF750 shut down. The RM750 eventually shut down when we increased that pulse to 106.25A.
    "at least an 850W" for 3090 or 3080?

    SF750 is currently the best SFX PSU in the market, do you think it would handle 10900K + 3090? maybe you already tested it?

  10. #9
    Join Date
    Oct 2006
    Location
    jonnyGURU forums, of course!
    Posts
    17,304
    Thanks Thanks Given 
    579
    Thanks Thanks Received 
    652
    Thanked in
    476 Posts

    Default

    Quote Originally Posted by Polaris20 View Post
    "at least an 850W" for 3090 or 3080?

    SF750 is currently the best SFX PSU in the market, do you think it would handle 10900K + 3090? maybe you already tested it?
    Without overclock.

  11. #10
    Join Date
    Aug 2018
    Location
    South Korea
    Posts
    518
    Thanks Thanks Given 
    203
    Thanks Thanks Received 
    65
    Thanked in
    41 Posts

    Default

    Quote Originally Posted by Jon Gerow View Post
    Without overclock.
    You mean you tested both overclock and stock
    and it worked fine on stock but nah no overclock?

    what program did you used to test it?

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •