What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

Comment Thread for AMD HD 7970 3GB Review

Sagath

Moderator
Staff member
Joined
Feb 7, 2009
Messages
3,925
Location
Edmonton, AB
Interestingly enough, PCIe 3.0 is garnering some results on other sites around the net, depending on workload. Results of up to 10% improvement have been seen in the AES encryption using the 7970 and PCIe 3 vs 2. May help with folding numbers too then since it seems GPU Compute is where the benefits will be seen, over gaming. :thumb:
 

JMCD

Well-known member
Joined
May 26, 2008
Messages
1,437
Location
London
Feels like a big jump like 4000>5000 wase, except with a bigger price tag. In any case, solid card and looks like they've got a lot of things right. Great review as always.
 

lKashl

Well-known member
Joined
Apr 26, 2009
Messages
191
Location
Australia
Nice and comprehensive as always SkYMTL, will you be giving us an update on the overclocking section once afterburner supports overvolting (or whichever other utility you prefer to use)? :D
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
13,410
Location
Montreal
Interestingly enough, PCIe 3.0 is garnering some results on other sites around the net, depending on workload. Results of up to 10% improvement have been seen in the AES encryption using the 7970 and PCIe 3 vs 2. May help with folding numbers too then since it seems GPU Compute is where the benefits will be seen, over gaming. :thumb:
The AES encryption benchmark a certain site used is notoriously inaccurate and gives highly conflicting results. I encourage you to run it and you'll see what I'm talking about. IMO, their result is a "ghost in the machine" rather than a repeatable aspect of performance. There's also a reason why there wasn't additional testing done by that site: PCI-E 3.0 brings zero benefits to this card at this time.

In addition, AES Encryption is a far cry away from OpenCL running [email protected] The algorithms are nowhere near the same.

Nice and comprehensive as always SkYMTL, will you be giving us an update on the overclocking section once afterburner supports overvolting (or whichever other utility you prefer to use)? :D
I do not overvolt ES cards. From past experience, it is like playing Russian Roulette, most of the time everything goes great but sometimes, the magic blue smoke appears.
 

Sagath

Moderator
Staff member
Joined
Feb 7, 2009
Messages
3,925
Location
Edmonton, AB
The AES encryption benchmark a certain site used is notoriously inaccurate and gives highly conflicting results. I encourage you to run it and you'll see what I'm talking about. IMO, their result is a "ghost in the machine" rather than a repeatable aspect of performance. There's also a reason why there wasn't additional testing done by that site: PCI-E 3.0 brings zero benefits to this card at this time.

In addition, AES Encryption is a far cry away from OpenCL running [email protected] The algorithms are nowhere near the same.
Well, good to know. Could have been nice to read about it in the article, at least a blurb?

As I stated: they don't.

Stanford's current GPU3 V3 core is a buggy mess that doesn't even recognize the card. I tried forcing GPU folding, changing priority, changing the GPUID, etc. Every time it dropped the GPU slot.
I assume he means when stanford fixes it :bleh:

Much like the rest of us, we are "patiently awaiting" folding results :thumb:
 

Latest posts

Twitter

Top