• Competitor rules

    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.

GPU prices go boom

Associate
Joined
1 Aug 2017
Posts
686
In an ideal world AMD, Intel and Nvidia would all be competing at the high end. It would be great for the consumer. I do think AMD can gain market share through their mobile business, plus switching to the 7nm process can only help too. It'll take time, but it can be done.

Also, if stock wasn't an issue, AMD should've had a price structure like the following: 480/580 - £150 / Vega 56 - £300 / Vega 64 - £400. For me this what the price of these GPU's should be (same goes for the Nvidia equivalents) rather than the price what they launched at, even before the mining/memory situation the cards were already overpriced.

If AMD are serious about gaining market share within the Desktop market, they need to price their GPU's realistically, rather than compete with Nvidia by pricing their GPU's at the same level.

As Anandtech once stated: There are no bad GPU's, just bad prices.
 
Permabanned
Joined
23 Jan 2014
Posts
465
Location
Beerwah, QLD Australia
In an ideal world AMD, Intel and Nvidia would all be competing at the high end. It would be great for the consumer. I do think AMD can gain market share through their mobile business, plus switching to the 7nm process can only help too. It'll take time, but it can be done.

Also, if stock wasn't an issue, AMD should've had a price structure like the following: 480/580 - £150 / Vega 56 - £300 / Vega 64 - £400. For me this what the price of these GPU's should be (same goes for the Nvidia equivalents) rather than the price what they launched at, even before the mining/memory situation the cards were already overpriced.

If AMD are serious about gaining market share within the Desktop market, they need to price their GPU's realistically, rather than compete with Nvidia by pricing their GPU's at the same level.

As Anandtech once stated: There are no bad GPU's, just bad prices.

HBM, it's more expensive, plus interposer assembly. AMD Vega was always going to be more expensive on a process basis.
 
Permabanned
Joined
23 Jan 2014
Posts
465
Location
Beerwah, QLD Australia
Regards the light bulb argument, there was a bulb in a New York Fire Dept station that lasted 70+ years because it was a low volt bulb and was never turned off. CBF Googling it but feel free if you like.
 
Soldato
Joined
27 Feb 2015
Posts
12,621
A card doing ETH mining is actually not been worked that hard, it can use less than half the power used when gaming, and can be run at around half the power limit depending on the binning of the card. So it does depend on what is been mined.
 
Associate
Joined
6 May 2011
Posts
1,467
Location
Milton Keynes
A card doing ETH mining is actually not been worked that hard, it can use less than half the power used when gaming, and can be run at around half the power limit depending on the binning of the card. So it does depend on what is been mined.

Pretty much this, a card that has been mining 24/7 will have not been subjected to multiple heating/cooling cycles, which is more detrimental to the PCB and components. Also as you point out most people will run their mining cards with lower voltages to keep power consumption down. I wouldn't be bothered about the actual GPU dying, but I would be worried about the fan.

I personally would buy an ex mining card if the price was right, but I don't mind the idea of replacing the cooler if it should fail.
 
Soldato
Joined
11 Jun 2003
Posts
10,795
Location
Hampshire
Well it kind of is, you posted an analogy, he pointed out the flaw in it, you got upset.

He ignored the point because it was valid, and pursued what he thought was a flaw because he didn’t have anything else to move forward with.

The point was regarding mean time to failure; nearly every mass produced product and component has an estimated mean time to, or between, failures.

I presented the bulb because I thought it would be simple enough to follow, evidently I should have used something even simpler. :p

Here’s a link that might be enlightening, shall we not derail any further now?

http://www.radio-electronics.com/in...-diodes/lifespan-lifetime-expectancy-mtbf.php
 
Soldato
Joined
26 May 2009
Posts
22,101
He ignored the point because it was valid
No he didn't, perhaps you didn't read the thread properly?

You tried to analogise the MTBF of graphics cards by using an example of a light left on for 5 hours a day and one running 24 hours a day, he posted to point out the flaw in the analogy (that a filament bulb will actually last longer running 24/7 than running 5 hours a day), you then replied that you didn't understand the difference.

The funniest thing is that the filament bulb example is actually quite an apt analogy for the GPU situation, where a mining card left running in moderate use will physically stand up better than a gaming card running higher usage in shorter bursts.
 
Soldato
Joined
30 Nov 2011
Posts
11,376
He ignored the point because it was valid, and pursued what he thought was a flaw because he didn’t have anything else to move forward with.

The point was regarding mean time to failure; nearly every mass produced product and component has an estimated mean time to, or between, failures.

I presented the bulb because I thought it would be simple enough to follow, evidently I should have used something even simpler. :p

Here’s a link that might be enlightening, shall we not derail any further now?

http://www.radio-electronics.com/in...-diodes/lifespan-lifetime-expectancy-mtbf.php

I think you need to read what you posted, it is quite apt - the biggest factors affecting LED and most electronics lifespan is temperature and being overdriven - e.g. gaming GPU being overclocked at 1100mv and run at 80C, vs mining GPU being run at 800mv and 60C
 
Back
Top Bottom