AMD Q2 2015 Financials Fall Short

[InvalidError]

When the "tons of people" were arguing for 4 GB 680s over the 2 GB models, actual published testing could find no significant difference in performance
When the "tons of people" were arguing for 4 GB 770s over the 2 GB models, actual published testing could find no significant difference in performance
When the "tons of people" were arguing for 4 GB 960s over the 2 GB models, actual published testing could find a significant difference in performance

When the GPU is already bottlenecked by memory bandwidth, adding more memory provides little to no performance benefits. This should not surprise anybody but bigger numbers still seem more futureproof. For only a few extra bucks, these often fall in the "why not" or "might as well" pricing range: better off spending $30 on an extra 2GB of VRAM that might never yield any performance benefits than regretting your $250 purchase for trying to shave $30 off the price.

 

[JackNaylorPE]

There's been no evidence that 4 GB is bottlenecking anything under normal game playing scenarios.... if you know of anything published indicating so, I'd love to read it.

That being said ..... it has nothing to do with the my post which but about public perception and marketing, both of which have little link to reality.

By putting out an 8 GB 390x and espousing it's "alleged benefits", the immediate inference by 99.7% of the buying public is that 4 GB is not enough ... thereby reinforcing the X GB is not enough mindset held by "tons of people ", ... a mindset held despite any published data indicating so.

 

[gigabob]

The interesting thing about AMD is that they compete on two fronts--GPUs and CPUs--with companies that each (for the most part, though not completely) focus on just one of those arenas: Nvidia in the graphics world and Intel in the CPU world. It takes enormous resources to double up like that, especially if a company expects to come out on top of the industry. Basically, they're spread too thin to win.

I use an AMD processor and and AMD graphics card and I'm happy with them. I like that they offer such cost effective options. The fact is, "bang for the buck/jack-of-all-trades" is their business strategy of choice and they will continue to lose ground until they make a true effort to be the best in the field, an effort that would require some big changes. Until then, people need to change their expectations. If you're looking for someone to challenge Nvidia or Intel, you'll have to look somewhere else.

 

[gigabob]

Intel would be dead if they were only on two front. Who ships more GPU's, by many times than their nearest competitor. When asked what distinguished AMD from other design houses, their flamboyant CEO replied, "Real companies have Fabs". With the sale of Global Foundries and debt held in Abu Dabi, Arab Micro Devices will be challanged to be a real player ever again. Especially if multi-gigabyte Storage extensions are added to desktops.

 

[jimmysmitty]

It will be interesting to see what happens in the CPU market with Intel being behind in the 7nm process. This will force both AMD and Intel to step up their games.

We have no idea where Intel is on the 7nm process and IBM has only shown SRAM which is a far cry from a working and viable CPU process node. There is still a very long time before we see a viable CPU running on 7nm.