AMD claims its GPUs beat Nvidia on performance per dollar
(2022/05/17)
- Reference: 1652816184
- News link: https://www.theregister.co.uk/2022/05/17/amd_gpu_nvidia/
- Source link:
As a slowdown in PC sales brings down prices for graphics cards, AMD is hoping to win over the market's remaining buyers with a bold, new claim that its latest Radeon cards provide better performance for the dollar than Nvidia's most recent GeForce cards.
In an [1]image tweeted Monday by AMD's top gaming executive, the chip designer claims its lineup of [2]Radeon RX 6000 cards provide better performance per dollar than competing ones from Nvidia, with all but two of the ten cards listed offering advantages in the double-digit percentages. AMD also claims to provide better performance for the power required by each card in all but two of the cards.
As a longtime gamer I'm grateful for the renewed competition in high-end graphics, we all win from it. As an [3]@AMD employee I'm super proud of what our [4]@Radeon team has accomplished. [5]#gamingpc [6]pic.twitter.com/6Rs9kjG9UD — Frank Azor (@AzorFrank) [7]May 16, 2022
There are a few important caveats to the claims. For one, AMD based its performance-per-dollar claim on two variables: the average frames per second (FPS) across nine games across multiple resolutions, and the lowest prices listed on NewEgg.com for AMD Radeon cards and Nvidia GeForce cards as of May 10.
This means mileage may vary depending on the game, the desired resolution of the game, and what graphics cards cost at any given time. Performance-per-dollar can be a helpful way to measure things, particularly for people on a budget, it's just that the variables can be moving targets, especially since AMD and Nvidia are making continuous optimizations with software.
To clarify, the table shows performance across multiple resolutions averaged out & not just at 4K. Thank you for sharing 4K performance, that delta is smaller and brings down the average overall number in the table. 1440p and 1080p performance tends to be even better. — Frank Azor (@AzorFrank) [8]May 17, 2022
We do wish that AMD would share the footnote referenced in the image that details how the comparisons were made so that anyone could properly review the claims. The Register only learned about AMD using nine games to determine the average FPS after asking a company spokesperson.
AMD's argument
With that out of the way, here are the takeaways from AMD's claims.
The chip designer says its flagship Radeon RX 6950 XT, with an $1,100 retail price, a 105 average FPS and 335 watts, provides 80 percent better performance per dollar and 22 percent better performance per watt than Nvidia's GeForce RTX 3090, which had a $1,700 retail price, a 91 average FPS and 350 watts. It's important to note that AMD doesn't have a comparable option for Nvidia's flagship, the [9]GeForce RTX 3090 Ti .
[10]
Further down the list, AMD says Radeon's performance-per-dollar advantage against Nvidia's competing GeForce cards ranged from six percent to 89 percent. Only two cards had advantages in the single-digit percentages while the rest were in the double-digit percentages.
[11]
[12]
As for the performance per watt of cards below the Radeon RX 6950 XT, AMD says seven out of the nine cards had advantages in the double-digit percentages. The other two, the Radeon 6750 XT and Radeon 6500 XT, didn't have performance per watt advantages.
[13]Nvidia open-sources Linux kernel GPU modules. Repeat, open-source GPU modules
[14]AMD approaches '30%' x86 CPU market share, thanks to servers 'n' laptops
[15]Intel blames software, COVID lockdowns for Arc GPU delays
[16]Nvidia pays $5.5m to settle SEC charges of shoddy cryptomining disclosures
When asked for comment, an Nvidia spokesperson sent The Register a link to [17]a Tuesday article from hardware test outfit Igor's Lab, which used measurement instruments to find that the power consumption of three Radeon 6000 XT cards is significantly higher than what software indicates.
Frank Azor, the AMD gaming executive who tweeted the Radeon claims, framed the chip designer's purported superiority against Nvidia as a reflection of the increasing competition in the GPU market, which Nvidia has dominated for years, in desktops and other segments, like servers.
The market will become even more competitive later this year when Intel manages to make its Arc GPUs widely available. The x86 giant last week [18]admitted that it is experiencing delays in the rollout of its [19]first gaming GPUs due to [20]COVID-19 lockdowns and "software readiness" issues, promising broader availability in the second half of the year. ®
Bootnote
AMD says the nine games used to determine the average FPS for each graphics card were the following: Assassin's Creed Valhalla, Borderlands 3, Doom Eternal, F1 2021, Forza Horizon 5, Horizon Zero Dawn, Far Cry 6, Shadow of the Tomb Raider, and Watch Dogs Legion.
Get our [21]Tech Resources
[1] https://twitter.com/AzorFrank/status/1526264896731680769
[2] https://www.theregister.com/2020/10/29/amds_radeon_gpus/
[3] https://twitter.com/AMD?ref_src=twsrc%5Etfw
[4] https://twitter.com/Radeon?ref_src=twsrc%5Etfw
[5] https://twitter.com/hashtag/gamingpc?src=hash&ref_src=twsrc%5Etfw
[6] https://t.co/6Rs9kjG9UD
[7] https://twitter.com/AzorFrank/status/1526264896731680769?ref_src=twsrc%5Etfw
[8] https://twitter.com/AzorFrank/status/1526569476757626880?ref_src=twsrc%5Etfw
[9] https://www.theregister.com/2022/03/29/nvidia_releases_1999_8kcapable_geforce/
[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[11] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[12] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[13] https://www.theregister.com/2022/05/12/nvidia_gpu_kernel/
[14] https://www.theregister.com/2022/05/13/amd_intel_x86/
[15] https://www.theregister.com/2022/05/11/intel_arc_delays/
[16] https://www.theregister.com/2022/05/06/nvidia_pays_55m_sec_fine/
[17] https://www.igorslab.de/en/graphics-cards-and-their-consumption-read-out-rather-than-measured-why-this-is-easy-with-nvidia-and-nearly-impossible-with-amd/4/
[18] https://www.theregister.com/2022/05/11/intel_arc_delays/
[19] https://www.theregister.com/2022/03/30/intel_arc_gpu/
[20] https://www.theregister.com/2022/04/28/semiconductor_firms_china_lockdowns_mess/
[21] https://whitepapers.theregister.com/
In an [1]image tweeted Monday by AMD's top gaming executive, the chip designer claims its lineup of [2]Radeon RX 6000 cards provide better performance per dollar than competing ones from Nvidia, with all but two of the ten cards listed offering advantages in the double-digit percentages. AMD also claims to provide better performance for the power required by each card in all but two of the cards.
As a longtime gamer I'm grateful for the renewed competition in high-end graphics, we all win from it. As an [3]@AMD employee I'm super proud of what our [4]@Radeon team has accomplished. [5]#gamingpc [6]pic.twitter.com/6Rs9kjG9UD — Frank Azor (@AzorFrank) [7]May 16, 2022
There are a few important caveats to the claims. For one, AMD based its performance-per-dollar claim on two variables: the average frames per second (FPS) across nine games across multiple resolutions, and the lowest prices listed on NewEgg.com for AMD Radeon cards and Nvidia GeForce cards as of May 10.
This means mileage may vary depending on the game, the desired resolution of the game, and what graphics cards cost at any given time. Performance-per-dollar can be a helpful way to measure things, particularly for people on a budget, it's just that the variables can be moving targets, especially since AMD and Nvidia are making continuous optimizations with software.
To clarify, the table shows performance across multiple resolutions averaged out & not just at 4K. Thank you for sharing 4K performance, that delta is smaller and brings down the average overall number in the table. 1440p and 1080p performance tends to be even better. — Frank Azor (@AzorFrank) [8]May 17, 2022
We do wish that AMD would share the footnote referenced in the image that details how the comparisons were made so that anyone could properly review the claims. The Register only learned about AMD using nine games to determine the average FPS after asking a company spokesperson.
AMD's argument
With that out of the way, here are the takeaways from AMD's claims.
The chip designer says its flagship Radeon RX 6950 XT, with an $1,100 retail price, a 105 average FPS and 335 watts, provides 80 percent better performance per dollar and 22 percent better performance per watt than Nvidia's GeForce RTX 3090, which had a $1,700 retail price, a 91 average FPS and 350 watts. It's important to note that AMD doesn't have a comparable option for Nvidia's flagship, the [9]GeForce RTX 3090 Ti .
[10]
Further down the list, AMD says Radeon's performance-per-dollar advantage against Nvidia's competing GeForce cards ranged from six percent to 89 percent. Only two cards had advantages in the single-digit percentages while the rest were in the double-digit percentages.
[11]
[12]
As for the performance per watt of cards below the Radeon RX 6950 XT, AMD says seven out of the nine cards had advantages in the double-digit percentages. The other two, the Radeon 6750 XT and Radeon 6500 XT, didn't have performance per watt advantages.
[13]Nvidia open-sources Linux kernel GPU modules. Repeat, open-source GPU modules
[14]AMD approaches '30%' x86 CPU market share, thanks to servers 'n' laptops
[15]Intel blames software, COVID lockdowns for Arc GPU delays
[16]Nvidia pays $5.5m to settle SEC charges of shoddy cryptomining disclosures
When asked for comment, an Nvidia spokesperson sent The Register a link to [17]a Tuesday article from hardware test outfit Igor's Lab, which used measurement instruments to find that the power consumption of three Radeon 6000 XT cards is significantly higher than what software indicates.
Frank Azor, the AMD gaming executive who tweeted the Radeon claims, framed the chip designer's purported superiority against Nvidia as a reflection of the increasing competition in the GPU market, which Nvidia has dominated for years, in desktops and other segments, like servers.
The market will become even more competitive later this year when Intel manages to make its Arc GPUs widely available. The x86 giant last week [18]admitted that it is experiencing delays in the rollout of its [19]first gaming GPUs due to [20]COVID-19 lockdowns and "software readiness" issues, promising broader availability in the second half of the year. ®
Bootnote
AMD says the nine games used to determine the average FPS for each graphics card were the following: Assassin's Creed Valhalla, Borderlands 3, Doom Eternal, F1 2021, Forza Horizon 5, Horizon Zero Dawn, Far Cry 6, Shadow of the Tomb Raider, and Watch Dogs Legion.
Get our [21]Tech Resources
[1] https://twitter.com/AzorFrank/status/1526264896731680769
[2] https://www.theregister.com/2020/10/29/amds_radeon_gpus/
[3] https://twitter.com/AMD?ref_src=twsrc%5Etfw
[4] https://twitter.com/Radeon?ref_src=twsrc%5Etfw
[5] https://twitter.com/hashtag/gamingpc?src=hash&ref_src=twsrc%5Etfw
[6] https://t.co/6Rs9kjG9UD
[7] https://twitter.com/AzorFrank/status/1526264896731680769?ref_src=twsrc%5Etfw
[8] https://twitter.com/AzorFrank/status/1526569476757626880?ref_src=twsrc%5Etfw
[9] https://www.theregister.com/2022/03/29/nvidia_releases_1999_8kcapable_geforce/
[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[11] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[12] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/personaltech&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YoQbB5-OEdnmrxiazVz7OwAAAFY&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[13] https://www.theregister.com/2022/05/12/nvidia_gpu_kernel/
[14] https://www.theregister.com/2022/05/13/amd_intel_x86/
[15] https://www.theregister.com/2022/05/11/intel_arc_delays/
[16] https://www.theregister.com/2022/05/06/nvidia_pays_55m_sec_fine/
[17] https://www.igorslab.de/en/graphics-cards-and-their-consumption-read-out-rather-than-measured-why-this-is-easy-with-nvidia-and-nearly-impossible-with-amd/4/
[18] https://www.theregister.com/2022/05/11/intel_arc_delays/
[19] https://www.theregister.com/2022/03/30/intel_arc_gpu/
[20] https://www.theregister.com/2022/04/28/semiconductor_firms_china_lockdowns_mess/
[21] https://whitepapers.theregister.com/
Annnnd can we
Do the same set of comparisons running on linux in wine? Anyone?