Why are AMD cpus prices being slashed to heavily?

[tential]

Thanks for reminding me. I forgot there was little to no R&D costs with GPUs.

There are R&D costs for gpus.
Despite years of evidence showing you there is more to pricing components other than pure manufacturing costs and die sizes, you still focus and harp in die size rather than look at the full picture of the cost to develop anything and bring it to market whether it is a cpu or gpu.

Anyway, you can continue with your sarcastic responses or but I think it would do you some good to look beyond die size when you think of the cost to develop something.


Either way, 1950x @ $700.... So good...

Buy 2 put them in the phantek case that allows for 2 mobos and enjoy 64 threads of Threadripper glory

 

[coercitiv]

Anyway, you can continue with your sarcastic responses or but I think it would do you some good to look beyond die size when you think of the cost to develop something.

Die size cost was the main argument invoked on this forum when some posters criticized Intel for not increasing core count sooner. We keep spinning in circles, alternatively invoking one argument over the other to justify pricing when it suits the narrative.

The truth is Intel's high profit margins have less to do with die size or R&D costs and more to do with the price for best performance the market is willing to pay for the chips. We just witnessed Intel increasing their core count by 50-100% for the same price in the mainstream space, effectively dropping their prices by 25-50% for the same MT performance, yet here we are talking about how AMD is forced to respond by adjusting prices.

 

[maddie]

Die size cost was the main argument invoked on this forum when some posters criticized Intel for not increasing core count sooner. We keep spinning in circles, alternatively invoking one argument over the other to justify pricing when it suits the narrative.

The truth is Intel's high profit margins have less to do with die size or R&D costs and more to do with the price for best performance the market is willing to pay for the chips. We just witnessed Intel increasing their core count by 50-100% for the same price in the mainstream space, effectively dropping their prices by 25-50% for the same MT performance, yet here we are talking about how AMD is forced to respond by adjusting prices.

True statement. AMD also, to a lesser degree.

I expect that we will see a relative price reduction over the next few years as INTEL and AMD battle for sales and that its possible for a full working 200mm^2 CPU die to sell for ~ $150, especially if sales volumes become more equalized, thus allowing R&D/die to be lowered. At least for AMD.

 

[scannall]

Intel has fought tooth and nail to keep CPU's from being considered a commodity. Which they could do, between Microsoft being x86 only and AMD with its construction cores being so bad. Well, it's a new landscape with AMD competitive now and CPU's are just a commodity. Which is great for US btw. Oh, and don't scream about R&D cost anymore either. The R&D into the Core architecture was paid for a long time ago, and there really isn't much of anything left to polish on it. Changing from a Ring Bus to a Mesh was a slight regression for the lower core count parts.

10nm should be interesting. Since by all indications it can't clock as high as the current 14nm. Say they manage to massage another 5% IPC, but lose 15% clockspeed, resulting in a net loss in ST performance. How loud will the howls of outrage be?

 

[VirtualLarry]

10nm should be interesting. Since by all indications it can't clock as high as the current 14nm. Say they manage to massage another 5% IPC, but lose 15% clockspeed, resulting in a net loss in ST performance. How loud will the howls of outrage be?

I expect Intel to limit those 10nm parts to mobile only, and release Kaby Lake Refresh 11! for desktop.

 

[tential]

Die size cost was the main argument invoked on this forum when some posters criticized Intel for not increasing core count sooner. We keep spinning in circles, alternatively invoking one argument over the other to justify pricing when it suits the narrative.

The truth is Intel's high profit margins have less to do with die size or R&D costs and more to do with the price for best performance the market is willing to pay for the chips. We just witnessed Intel increasing their core count by 50-100% for the same price in the mainstream space, effectively dropping their prices by 25-50% for the same MT performance, yet here we are talking about how AMD is forced to respond by adjusting prices.

You've explained this far better than I did. Thanks