Caporegime
I see you went for realism. UK & heavy rain
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
I see you went for realism. UK & heavy rain
Is there a version of gpu-z or something else that shows wattage?
I don't think any of the AOIs are hitting 1750 since the 17.8.2 drivers, mine is also topping out at around 1720. If you look back a few pages AMD Matt explains why this is, apparently it's normal behaviour so I'd hold fire on the RMA. Also, if you look at all the listings for the AOI cards, it has the boost clock down as 1677 and not 1750 so unless you can't boost past the listed 1677 I'd say there is nothing wrong with your card.So I got my new Corsair HX1000i installed after work last night and spent most of the evening testing.
With the PSU ruled out I can confirm that my card IS unstable at stock (Balanced mode or higher) clocks. It's not by a lot though. It's clear that the stock max clock of 1750Mhz is simply more than the actual silicon is capable of. The hard limit for mine seems to be around 1720Mhz.
So poor binning and inadequate testing seems to be the culprit here.
Slightly undervolted and max core clock set at 1700Mhz, it appears to be completely stable. I need to do some longer duration stress tests to prove, but it looped heaven and firestrike each for 45 minutes without crashing. Thermals in particular, but also power usage are much nicer here.
In all other respects the card is excellent - HBM2 clocks nicely, 1100Mhz at stock volts is a breeze, pump and fan are very quiet, no coil whine until it hits silly fps.
So I'm weighing the pros and cons of RMA now. Seems like a bit of an overreaction to return because of the last 50Mhz of cpu core that I would be unlikely to ever use outside of benching, especially considering what little difference that makes compared to memory OC on these. Gonna mull it over for a few days before I do anything.
Good news!New card definitely approved on my end, runs benchmarks just fine and more importantly games without any black screen crashes like the previous one. I'm now certain that my previous Sapphire card was faulty.
I'm now a very happy Vega owner
Aldo wanted to mention that for some reason this card is quieter than the previous on load, definitely seems like the fan on the radiator is not spinning as high as the previous one.
You may be right but the conclusion I have come to from following this thread from the start is that if the card is working correctly then a 850w PSU will be fine. I haven't yet heard of anyone having card problems with an 850 PSU to then replace it with a 1000 PSU and their problems have gone away, it's always turned out to be the card unless I've missed something...Whether you can use a lesser psu seems to be on a card by card basis, I think it's why Vega had a delay and why we have Revision 0's in the AIO's and Revision 1's in the air-cooled models. I tried two different 850 watt psu's with my AIO neither up to the task. So while there is a lot of faulty cards we can't say an 850watt psu is more than enough for an AIO.
I hope the AIB models (ie only the Strix so far) all use Revision 1 chips.
That's not the issue he's having, his card crashes at default settings which is BS. That's why several of us returned our cards for new ones. AMDMatt shows the card is definitely capable of hitting 1750, he just clocked his differently in order to sustain higher clocks on 17.8.2 since it seems to handle boost a bit differently than previous drivers.I don't think any of the AOIs are hitting 1750 since the 17.8.2 drivers, mine is also topping out at around 1720. If you look back a few pages AMD Matt explains why this is, apparently it's normal behaviour so I'd hold fire on the RMA. Also, if you look at all the listings for the AOI cards, it has the boost clock down as 1677 and not 1750 so unless you can't boost past the listed 1677 I'd say there is nothing wrong with your card.
That's not the issue he's having, his card crashes at default settings which is BS. That's why several of us returned our cards for new ones. AMDMatt shows the card is definitely capable of hitting 1750, he just clocked his differently in order to sustain higher clocks on 17.8.2 since it seems to handle boost a bit differently than previous bios. Also, Sapphire advertised the card as having DP7 at 1750 when the announced the card, you can see it here about halfway down the page. https://sapphirenation.net/sapphire-radeon-vega/
Ah ok I didn't realise he was getting crashes, just that he couldn't get the card to boost up to 1750 which I haven't been able to do since the 17.8.2 drivers. Where did Matt show the card boosting to 1750 on the 17.8.2 drivers? Has anyone seen there card actually hit 1750 since updating to the latest drivers?That's not the issue he's having, his card crashes at default settings which is BS. That's why several of us returned our cards for new ones. AMDMatt shows the card is definitely capable of hitting 1750, he just clocked his differently in order to sustain higher clocks on 17.8.2 since it seems to handle boost a bit differently than previous drivers.
Also, Sapphire advertised the card as having DPM7 at 1750 when they announced the card, you can see it here about halfway down the page. https://sapphirenation.net/sapphire-radeon-vega/
I wouldn't trust GPU-Z 2.2.0 TBH. I know it says revision C1 but a lot of the other information regarding Vega is missing as are half the sensor readings. It even says the HBM2 memory is by Micron! Also if you click on lookup it tells you it can't find this device in its database.
Ah ok I didn't realise he was getting crashes, just that he couldn't get the card to boost up to 1750 which I haven't been able to do since the 17.8.2 drivers. Where did Matt show the card boosting to 1750 on the 17.8.2 drivers? Has anyone seen there card actually hit 1750 since updating to the latest drivers?
New card definitely approved on my end, runs benchmarks just fine and more importantly games without any black screen crashes like the previous one. I'm now certain that my previous Sapphire card was faulty.
I'm now a very happy Vega owner
Aldo wanted to mention that for some reason this card is quieter than the previous on load, definitely seems like the fan on the radiator is not spinning as high as the previous one.
Check this out, this is with the Wattman settings linked above.Ah ok I didn't realise he was getting crashes, just that he couldn't get the card to boost up to 1750 which I haven't been able to do since the 17.8.2 drivers. Where did Matt show the card boosting to 1750 on the 17.8.2 drivers? Has anyone seen there card actually hit 1750 since updating to the latest drivers?
I stand corrected. Thanks for that. I actually matched Matt's settings on that video but Wattman is showing my max clock as 1724 when gaming or benching...
It can vary a bit from game to game, so it may not always be as high as what i have in Dirt 4 above.I stand corrected. Thanks for that. I actually matched Matt's settings on that video but Wattman is showing my max clock as 1724 when gaming or benching...
Do you have a pic please.