Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
So is anyone actually running a GTX 1070/80?
- Thread starter Messy
- Start date
- Sort by reaction score
Check the date of this article: it was released BEFORE whole situation got any traction. Information came nowhere else, but from Nvidia.
Their information about GP100 bump gate is also accurate. Because it also came from two sources, that are most interested: TSMC and Nvidia.
Their information about GP100 bump gate is also accurate. Because it also came from two sources, that are most interested: TSMC and Nvidia.
You should have seen Charlie's rant after R9 3XX series "refresh". It was hilarious how he mocked AMD for it.
So you completely ignore that they have informed about this issue, and pointed out the reasons behind this issue, BEFORE it got traction of forums, and even end users found it?
All I can tell you, because I would break the policy of SA site is that this issue affects only GDDR5X memory GPUs, and it has everything to do with Memory Controller and Memory itself.
GTX 1070 and 1060 are free from this issues. Pascal Titan X is not, and will get affected.
I'm not going to get too worried about the issue until I either 1) See it reported on sites like anandtech or ars technica or 2) notice it on my own gtx 1080. Until then I am very happy with my card.
All Anandtech will do is report that Nvidia is working on driver fix, and we should expect it "soon".
And driver will not fix the problem.
Anyways, enjoy your GTX 1080. Glad that you have one working.
Yeah, but once there is an anandtech article on it goes from random people on a message board talking about it to a reputable news organization who has likely communicated with nvidia about it.
I can't wait for Nvidia to come out with their pascal drivers so I can turn it into a kickass hackintosh
You're entitled to believe whatever you like I'm not going to argue with you.
http://www.tweaktown.com/news/53375/nvidias-next-gen-gpus-made-14nm-samsung/index.html
Somebody has been arguing with me that I am reading too much in some information. Volta is coming next year, and it will be made on Samsung process.
Somebody has been arguing with me that I am reading too much in some information. Volta is coming next year, and it will be made on Samsung process.
A hermit in a cave would know that Volta will be Nvidia's GTC 2017 headline next spring/early summer.
ATI stock is going to drop.
That last bit is more unlikely than Trump winning the vote in United States.
i'm selling a EVGA 1080 GTX with a dual mini 6-pin to 8-pin made specially for the Mac Pro so that your 1080 GTX will get full power from your Mac Pro and if your guys interested go on Amazon user name 34Skyline under EVGA 1080 GTX founder edition
There's a really good post on the HardOCP forums about "asynchronous compute", particularly how AMD and NVIDIA implement it:
https://hardforum.com/threads/demystifying-asynchronous-compute-v1-0.1909504/
Should clear up any confusion, especially relating to how NVIDIA implements this (i.e. it's not just preemption as some have claimed here in the past).
https://hardforum.com/threads/demystifying-asynchronous-compute-v1-0.1909504/
Should clear up any confusion, especially relating to how NVIDIA implements this (i.e. it's not just preemption as some have claimed here in the past).
Thanks for sharing this in many different threads across several different forums.
