AMD boss confirms high-end Navi cards are "on track"

3) Go look at more benchmarks because your claims of the RX5700 not being power efficient and too hot makes no sense. It only runs hot because of its blower fan design, not because the chip itself is power hungry.

The 5700 and 5700 XT have very comparable power for performance ratio as the RTX2000s. Look at these benchmarks charts of Performance Per Watt from Techpowerup:

Being that the 5700XT is 251mm and 2060 Super is 445mm in die size, while they have the same TDP, 5700XT has way higher thermal density that makes it harder to cool. Look at the temperature difference between stock Vega 56 and 5700XT, these two have the same TDP (even the 180W TDP 5700 is hotter than Vega 56).

https://www.techpowerup.com/review/amd-radeon-rx-vega-56/36.html
https://www.techpowerup.com/review/amd-radeon-rx-5700-xt/32.html

So yeah AIB 5700XT can bring the noise down compared to stock but the temp will remain quite high, I think applying liquid metal here will help tremendously for anyone with the balls to do that (I put Conductonaut on my watercooled 1080 Ti and 2080 Ti and they have been working quite well).
 
Last edited:
Hey AMD fan. It’s obviously a hot card, get over it. You can throw all sorts of graphs at me and tell me I can’t compare a card because of a price difference. The fact is, it’s already a very hot card and if you make it bigger it will be hotter still. Practically every reviewer states it’s a hot card and would benefit hugely from watercooling. Also you should see the power consumption when overclocked, it’s horrific!

You seem to be desperate from preventing the truth that this card runs hot from getting out there. Why? Do you own AMD stock?

*cough*
He totally blew Your bias drivel right out of the water, and that^ is your come back. Your defense?

Both cards measure temps differently, both card thermals are different. You are skewing facts and spreading myths.
 
Being that the 5700XT is 251mm and 2060 Super is 445mm in die size, while they have the same TDP, 5700XT has way higher thermal density that makes it harder to cool. Look at the temperature difference between stock Vega 56 and 5700XT, these two have the same TDP (even the 180W TDP 5700 is hotter than Vega 56).

https://www.techpowerup.com/review/amd-radeon-rx-vega-56/36.html
https://www.techpowerup.com/review/amd-radeon-rx-5700-xt/32.html

So yeah AIB 5700XT can bring the noise down compared to stock but the temp will remain quite high, I think applying liquid metal here will help tremendously for anyone with the balls to do that (I put Conductonaut on my watercooled 1080 Ti and 2080 Ti and they have been working quite well).

The smaller die will not be a big problem because we have a trickle of videos and reviews that show switching the cooler or even just increasing the fanspeed on the stock blower cooler will make a huge difference and keep the temperatures down. This isn't like the situation when Intel used garbage thermal paste with the Ivy Bridge and Haswell processors that required delidding and repasting with liquid metal.

According to Hardware Unbocked's tests, the stock drivers only have the stock blower fan speed at 33% of the max at around 1900 RPM. Hardware Unbocked did a test where they cranked up the fanspeed to 50% fanspeed at 2700 RPM which increased the max core frequency and lowered the gaming temperature to 69-70'C. So by simply cranking up the stock blower cooler's fanspeed curve to 50%, they lowered the 5700 XT's gaming temperature from 85'C to 70'C.
Link:

JayZeeTwoCents also did a test with a waterblock on the 5700 XT and it maxed out in the low 60s'C with a +30watt increase & a higher sustained boost speed, and ran in the mid 60s'C with a +50watt increase & overclock. So these 5700 XT will be significantly cooler with the new AIB coolers with better heatsinks/fans and with watercooling. (though watercooling isn't really worth it according to him due to the price/perf ratio)
 
The smaller die will not be a big problem because we have a trickle of videos and reviews that show switching the cooler or even just increasing the fanspeed on the stock blower cooler will make a huge difference and keep the temperatures down. This isn't like the situation when Intel used garbage thermal paste with the Ivy Bridge and Haswell processors that required delidding and repasting with liquid metal.

According to Hardware Unbocked's tests, the stock drivers only have the stock blower fan speed at 33% of the max at around 1900 RPM. Hardware Unbocked did a test where they cranked up the fanspeed to 50% fanspeed at 2700 RPM which increased the max core frequency and lowered the gaming temperature to 69-70'C. So by simply cranking up the stock blower cooler's fanspeed curve to 50%, they lowered the 5700 XT's gaming temperature from 85'C to 70'C.
Link:

JayZeeTwoCents also did a test with a waterblock on the 5700 XT and it maxed out in the low 60s'C with a +30watt increase & a higher sustained boost speed, and ran in the mid 60s'C with a +50watt increase & overclock. So these 5700 XT will be significantly cooler with the new AIB coolers with better heatsinks/fans and with watercooling. (though watercooling isn't really worth it according to him due to the price/perf ratio)

You didn't listen carefully then, on the 5700XT Steve crank the fan to 100% (4600rpm) and the temp was ~70C, stock temp was 88C with 43% fan and this is still winter in Australia where ambient temp is 20C, F1 isn't even an intensive game also. On the normal 5700 Steve cranked the fan to 50% and temp were 68C.
Yes with better cooler you can reduce the temp of the 5700XT but don't expect the any cheap AIB model to improve much over stock. Overall any 5700XT model will be hotter and louder than the comparative 2060 Super model simple because of the higher thermal density of the 5700XT.
 
I'm glad AMD exists, truly I am, but I have genuinely never been impressed by any of their products. Not one.
 
You didn't listen carefully then, on the 5700XT Steve crank the fan to 100% (4600rpm) and the temp was ~70C, stock temp was 88C with 43% fan and this is still winter in Australia where ambient temp is 20C, F1 isn't even an intensive game also. On the normal 5700 Steve cranked the fan to 50% and temp were 68C.
Yes with better cooler you can reduce the temp of the 5700XT but don't expect the any cheap AIB model to improve much over stock. Overall any 5700XT model will be hotter and louder than the comparative 2060 Super model simple because of the higher thermal density of the 5700XT.

No, there are two different scenarios where he mentions 70'C. I was talking about the second scenario. The first scenario is with the 100% fan curve where he was overclocking with the 5700XT so its power consumption reached over around 250 watts. That's where Steve set the fan to 100% and the overclocked 250w GPU reaches 70'C. The second scenario is around the video's 12:15 mark where he keeps the 5700XT at stock speeds and ~160W, where the GPU maxes out at 70'C with a 50% fancurve. For the second scenario, you see 68-69'C on the screen, but Steve specifically says it reaches 70'C.

Here is the video time for the second scenario where he uses the non-overclocked 160W 5700XT at 50% fancurve to hit 70'C. He specifically says 70'C at 12:33.

As for your statement that the game isn't very intensive, as long as the framerates are not artificially capped and there is no bottleneck from other components such as the CPU, the GPU should still hit 100% core useage. It's no furmark, but should still provide a decent representation of what typical gaming should be like.

The 5700XT can likely reach temperatures and noise level comparable to the 2070 and 2060S with decent AIB coolers. For comparison, the 5700 XT is 84'C out of the box with the lackluster blower cooler. According to Tech2060 Super Founder's Edition reaches 76'C in gaming with a non-blower cooler with 2 fans. Any 5700XT with a half decent AIB cooler can easily drop the temperature by 8'C and hit 2060 Super FE's temperatures.

How do I know? Here is a comparison of the 1070 FE with the blower cooler vs the 1070 with an MSI dual fan cooler. The 1070 FE blower cooler hits 83'C while the 1070 with the MSI dual fan hits only 71'C, a drop of 12'C just by changing the cooler from the blower model to a regular 2 fan model. The noise also drops from the 36 dBa of the 1070 FE to 29 dBa of the 1070 MSI model.

https://www.techpowerup.com/review/nvidia-geforce-gtx-1070/28.html
https://www.techpowerup.com/review/msi-gtx-1070-gaming-z/30.html

2060S FE temps:
https://www.techpowerup.com/review/msi-geforce-rtx-2060-super-gaming-x/31.html
 
Last edited:
I'm glad AMD exists, truly I am, but I have genuinely never been impressed by any of their products. Not one.

What about when the AMD 5000 series came out? Nvidia's Fermi 400 counterparts ran hotter, used more power for the same or less performance, and the top tier Nvidia models were actually slower than AMD's top tier models.

The GTX480 was 20% slower than an HD 5970 but used 50 watts more electricity. In terms of performance per watt efficiency, it was 60% worse compared to the HD5870, 50% worse compared to the HD5970, and was around 90% worse compared to the HD5850.
 
What about when the AMD 5000 series came out? Nvidia's Fermi 400 counterparts ran hotter, used more power for the same or less performance, and the top tier Nvidia models were actually slower than AMD's top tier models.

The GTX480 was 20% slower than an HD 5970 but used 50 watts more electricity. In terms of performance per watt efficiency, it was 60% worse compared to the HD5870, 50% worse compared to the HD5970, and was around 90% worse compared to the HD5850.
You're talking about hardware I never owned. As a typical consumer, I have not actually owned a ton and have only built a total of 5 systems of my own.

With only 2 choices it's easy to get stuck into seeming or even feeling like a fan boy, but I always looked objectively at what was available at any given time and for my budget and usage, Intel and nvidia were always the more logical choice at the times I was ready to spend.
 
No, there are two different scenarios where he mentions 70'C. I was talking about the second scenario. The first scenario is with the 100% fan curve where he was overclocking with the 5700XT so its power consumption reached over around 250 watts. That's where Steve set the fan to 100% and the overclocked 250w GPU reaches 70'C. The second scenario is around the video's 12:15 mark where he keeps the 5700XT at stock speeds and ~160W, where the GPU maxes out at 70'C with a 50% fancurve. For the second scenario, you see 68-69'C on the screen, but Steve specifically says it reaches 70'C.

Here is the video time for the second scenario where he uses the non-overclocked 160W 5700XT at 50% fancurve to hit 70'C. He specifically says 70'C at 12:33.

As for your statement that the game isn't very intensive, as long as the framerates are not artificially capped and there is no bottleneck from other components such as the CPU, the GPU should still hit 100% core useage. It's no furmark, but should still provide a decent representation of what typical gaming should be like.

The 5700XT can likely reach temperatures and noise level comparable to the 2070 and 2060S with decent AIB coolers. For comparison, the 5700 XT is 84'C out of the box with the lackluster blower cooler. According to Tech2060 Super Founder's Edition reaches 76'C in gaming with a non-blower cooler with 2 fans. Any 5700XT with a half decent AIB cooler can easily drop the temperature by 8'C and hit 2060 Super FE's temperatures.

How do I know? Here is a comparison of the 1070 FE with the blower cooler vs the 1070 with an MSI dual fan cooler. The 1070 FE blower cooler hits 83'C while the 1070 with the MSI dual fan hits only 71'C, a drop of 12'C just by changing the cooler from the blower model to a regular 2 fan model. The noise also drops from the 36 dBa of the 1070 FE to 29 dBa of the 1070 MSI model.

https://www.techpowerup.com/review/nvidia-geforce-gtx-1070/28.html
https://www.techpowerup.com/review/msi-gtx-1070-gaming-z/30.html

2060S FE temps:
https://www.techpowerup.com/review/msi-geforce-rtx-2060-super-gaming-x/31.html

Okay use auto subtitle on Youtube if you can't listen to English properly, "as for the standard 5700...", Steve was talking about the 5700 in your second scenario. The result for 5700 XT were:
84C stock 88C overclocked with 43% fan speed and 70C overclocked with 100% fan speed
Result for 5700:
74C stock 85C overclock with 33% fan speed and 68C overclocked with 50% fan speed.
These results are horrible considering it's winter in Australia where ambient temp is below 20C, imagine these cards in 30C ambient.


I'm glad AMD exists, truly I am, but I have genuinely never been impressed by any of their products. Not one.

What about when the AMD 5000 series came out? Nvidia's Fermi 400 counterparts ran hotter, used more power for the same or less performance, and the top tier Nvidia models were actually slower than AMD's top tier models.

The GTX480 was 20% slower than an HD 5970 but used 50 watts more electricity. In terms of performance per watt efficiency, it was 60% worse compared to the HD5870, 50% worse compared to the HD5970, and was around 90% worse compared to the HD5850.

Well not to insult you but you need to check your facts properly, the HD 5970 is a dual GPU card and crossfire was never a good solution, not now and definitely not back then, the FPS counter is very high but it's micro-shuttering all over the place (thanks to Techreport's "inside the second" article debunking the usefulness of crossfire).
 
Okay use auto subtitle on Youtube if you can't listen to English properly, "as for the standard 5700...", Steve was talking about the 5700 in your second scenario. The result for 5700 XT were:
84C stock 88C overclocked with 43% fan speed and 70C overclocked with 100% fan speed
Result for 5700:
74C stock 85C overclock with 33% fan speed and 68C overclocked with 50% fan speed.

Yes, that is correct. Of course the temperatures are not "amazing" because the fancurve is weak and the stock blower fan is rubbish. But it's not nearly as bad as you make it seem.

It's running at 84'C with a low fan speed with a blower cooler. If you look at previous GPUs with the blower model - the 1070, for instance, it also ran at around 83' Celsius from Techpowerup's test and 79 or 80'C from Techspot's test.

This Techspot video established that turning up the fan speed will do wonders for the cooling. The other video I linked earlier shows water cooling getting it down to 60s'C. I imagine a half decent AIB will be somewhere in the middle between 60s'C and 80s'C. No need for liquid metal.

These results are horrible considering it's winter in Australia where ambient temp is below 20C, imagine these cards in 30C ambient.

1) I'm pretty sure reviewers account for ambient temperatures. They're not going to test one card at 20'C and another card at 30'C.

Go look at Techspot's review of the 1070 - it also hit about 79'C. Assuming Techspot tests their computers at comparable ambient, that is about a 5'C difference between the 1070 and the 5700XT.

2) Also, I'm pretty sure people in Australia have air conditioning. Who runs a gaming computer in a warm room in the mid-high 80s'F? A more realistic room temperature would be something in the mid 70s'F or 24'C.

Well not to insult you but you need to check your facts properly, the HD 5970 is a dual GPU card and crossfire was never a good solution, not now and definitely not back then, the FPS counter is very high but it's micro-shuttering all over the place (thanks to Techreport's "inside the second" article debunking the usefulness of crossfire).

I know the 5970 is a two GPU card. I never claimed it was a single GPU card in the first place.

It doesn't change the fact that a dual GPU card from AMD got higher framerates while consuming less power than Nvidia's GTX480. It was still a single card that was the performance king, and SLi/Xfire and dual cards were all the rage back then with Nvidia releasing the GTX590 later on.

The first release GTX400 series was basically Nvidia's version of Bulldozer - power inefficient and hot. It was a good time to favor AMD over Nividia, and the GTX480 became the butt of jokes for years for how hot and noisy it got.
 
Yes, that is correct. Of course the temperatures are not "amazing" because the fancurve is weak and the stock blower fan is rubbish. But it's not nearly as bad as you make it seem.

It's running at 84'C with a low fan speed with a blower cooler. If you look at previous GPUs with the blower model - the 1070, for instance, it also ran at around 83' Celsius from Techpowerup's test and 79 or 80'C from Techspot's test.

This Techspot video established that turning up the fan speed will do wonders for the cooling. The other video I linked earlier shows water cooling getting it down to 60s'C. I imagine a half decent AIB will be somewhere in the middle between 60s'C and 80s'C. No need for liquid metal.

1) I'm pretty sure reviewers account for ambient temperatures. They're not going to test one card at 20'C and another card at 30'C.

Go look at Techspot's review of the 1070 - it also hit about 79'C. Assuming Techspot tests their computers at comparable ambient, that is about a 5'C difference between the 1070 and the 5700XT.

2) Also, I'm pretty sure people in Australia have air conditioning. Who runs a gaming computer in a warm room in the mid-high 80s'F? A more realistic room temperature would be something in the mid 70s'F or 24'C.

I know the 5970 is a two GPU card. I never claimed it was a single GPU card in the first place.

It doesn't change the fact that a dual GPU card from AMD got higher framerates while consuming less power than Nvidia's GTX480. It was still a single card that was the performance king, and SLi/Xfire and dual cards were all the rage back then with Nvidia releasing the GTX590 later on.

The first release GTX400 series was basically Nvidia's version of Bulldozer - power inefficient and hot. It was a good time to favor AMD over Nividia, and the GTX480 became the butt of jokes for years for how hot and noisy it got.

Well we have another great example of how hot 7nm is, that is the Radeon VII, even with the triple fan cooler that card is still hot and loud. I think with bigger Navi chips it's best to go watercooling, either using AIO or custom. The price/perf of the 5700 XT is a bit misleading because the the stock cooler is inadequate, basically you have to buy more expensive AIB card or replace the cooler to get comparable noise profile to Nvidia 2060 Super or 2070 Super.

I live in a warm climate country so it's always above 30C every day of the year, I keep my AC at 27C (80F) to save electricity and to avoid thermal shock. Needless to say AMD is not very popular in warm countries.

I know ATI/AMD produced great graphic cards, the GPU battle were even years ago but right now AMD kinda ruined it. Last noteworthy AMD GPU I bought were the R9 290, slapped on the Arctic Accelero 4 for another 80usd and rocked it for awhile until AMD driver basically killed it by locking the fan to 20%. So yeah I had some very minor annoyance with Nvidia driver time to time but never a hardware killing one lol.

GPU I have owned: 5200 Ultra, 6800 GT, 7800 GT, 7950 GX2, 8800 GT, ATI HD 4890, GTX 460, GTX 680, AMD R9 290, GTX 980, GTX Titan X, GTX 1080 Ti and RTX 2080 Ti. Out of those I only regret buying the 5200 Ultra and the 7800 GT because the ATI had better cards for the same price. Oh yeah and the 8800 GT kept dying one by one lol, I had 4 EVGA 8800 GT replacements and they all died within 2 months.
 
Last edited:
Well we have another great example of how hot 7nm is, that is the Radeon VII, even with the triple fan cooler that card is still hot and loud. I think with bigger Navi chips it's best to go watercooling, either using AIO or custom. The price/perf of the 5700 XT is a bit misleading because the the stock cooler is inadequate, basically you have to buy more expensive AIB card or replace the cooler to get comparable noise profile to Nvidia 2060 Super or 2070 Super.

The Radeon VII is a different architecture and peaks at over 300watts in gaming according to the benchmark reviews. So it's going to run quite a bit hotter than the new AMD RDNA cards when the the 5700XT peaks at around 220s watt and the 5700 which peaks at around 180watts. I think most models could get away with 2 decent fans, while the 3 fan versions would be fairly quiet compared to the VII.

The fact that the 5700XT performs about the same as the Radeon VII while consuming 80 watts less is a decent step for AMD's performance efficiency.

As for AIB, they're not necessarily all going to be more expensive. I've found some AIB GPUs with comparable prices as the stock models during past releases (and with decent cooling with at least 2 fans too that were better than the 1 blower fan model).

I live in a warm climate country so it's always above 30C every day of the year, I keep my AC at 27C (80F) to save electricity and to avoid thermal shock. Needless to say AMD is not very popular in warm countries.

Yeh, that is understandable. Nvidia has won the performance per watt crown since the GTX900 Maxwells.

I know ATI/AMD produced great graphic cards, the GPU battle were even years ago but right now AMD kinda ruined it. Last noteworthy AMD GPU I bought were the R9 290, slapped on the Arctic Accelero 4 for another 80usd and rocked it for awhile until AMD driver basically killed it by locking the fan to 20%. So yeah I had some very minor annoyance with Nvidia driver time to time but never a hardware killing one lol.

GPU I have owned: 5200 Ultra, 6800 GT, 7800 GT, 7950 GX2, 8800 GT, ATI HD 4890, GTX 460, GTX 680, AMD R9 290, GTX 980, GTX Titan X, GTX 1080 Ti and RTX 2080 Ti. Out of those I only regret buying the 5200 Ultra and the 7800 GT because the ATI had better cards for the same price. Oh yeah and the 8800 GT kept dying one by one lol, I had 4 EVGA 8800 GT replacements and they all died within 2 months.

My first card was the FX 5600 for the regular PCI slot IIRC. Went with the 9800GTX+ as my second card, and then all Nvidia until the AMD's 5000 and 7000 series, then back to Nvidia with the 1000 series.

I'd say AMD is making a comeback with these new RDNA cards. The cards themselves have comparable power for performance efficiency to Pascal and Turing GPUs. AMD's Polaris cards were behind Nvidia's Pascal and only achieved Maxwell's efficiency, while these RDNAs fall between Pascal and Turing. They need to fix the cooler/heatsink and it'll be a decent rival to Nvidia's lineup.

As for the R9 290s, those R7/9 cards were the refreshed GCN 2.0 7000 cards, and IIRC were the last time they kept up with Nvidia in performance per watt efficiency until now.
 
I'm glad AMD exists, truly I am, but I have genuinely never been impressed by any of their products. Not one.

You must be really young. About 12 years ago AMD had much faster CPUs than Intel. While Intel had higher clocks, AMD had a more efficient superscalar architecture, with better IPC. That means it was executing more instructions per clock. The results were spectacular. For example a 1.6 GHz AMD chip was outperforming a 3.2 GHz Intel chip. AMD was consistently making better Intel-compatible chips than Intel themselves.

Same goes for Radeon cards. When they were produced by ATI (which is the same company that is now part of AMD) their Radeon 9700/9800 cards were sweeping the floor with Nvidia cards. Nvidia was crying in the media, attacking benchmark developers, faking benchmark results, just to come close to Radeon. And to be honest, Nvidia was always cheating in benchmarks, even today. They have special deals with game makers, bribing them to optimize games to run normal on Nvidia, but slower on AMD.

When AMD chips had more ROPs and faster Mem, but less Cores, Nvidia would bribe the game producers to use shaders that require more Cores, without being heavy on ROPs or Mem. When AMD increased the number of Cores, but was behind on ROPs, Nvidia would bribe their "friendly" game devs to utilize ROPs, but not CUs.

That's why in the games with Nvidia logo AMD cards are unbelievably slow. It's more politics than technology.
 
I hope not.
I hope they stay away from the top end because I enjoy the fact the nvidia fan boys get to dig deep to purchase their cards and get to criticise AMD for not delivering competition.
No No, let AMD release a faster GPU than the 2080ti. AMD will always dominate in Price to Performance. This would help squeeze Nvidia's bottom line as much as possible. Especially after there scam GPP nonsense.
 
1) First, running hotter is not the same as using more power/producing more heat. A fanless RX460 can run hotter than an RX590 with 3 cooling fans even though the later has 3 times to TDP. Your comments about the RX5700s running hot like a space heater preventing more cores in the higher variants implies you don't realize the RX5700 actually has similar overall power consumption and heat output per performance as the RTX series.

2) Comparing a 5700/5700XT with a 2070 Super makes no sense when the 2070 Super is $100 more expensive. The RX5700 XT is $400 while the 2070 Super is $500. The 5700 XT is at the price point of the regular RTX2070 or 2060 Super and it performs roughly comparable to the RTX2070.

In terms of power for performance, the RX5700XT is slightly less efficient than the 2070 Regular by only 10 watts, while the RX5700 is more power efficent beats the 2060 regular pretty easily while consuming less power in Techspot's benchmarks.

Power2.png


https://www.techspot.com/review/1870-amd-radeon-rx-5700/


3) Sp again, you need to go look at the benchmarks because your claims of the RX5700 not being power efficient and too hot makes no sense. It only runs hot because of its blower fan design, not because the chip itself is power hungry.

The 5700 and 5700 XT have very comparable power for performance ratio as the RTX2000s. Look at these benchmarks charts of Performance Per Watt from Techpowerup:


performance-per-watt_1920-1080.png

performance-per-watt_1920-1080.png
Hey AMD fan. It’s obviously a hot card, get over it. You can throw all sorts of graphs at me and tell me I can’t compare a card because of a price difference. The fact is, it’s already a very hot card and if you make it bigger it will be hotter still. Practically every reviewer states it’s a hot card and would benefit hugely from watercooling. Also you should see the power consumption when overclocked, it’s horrific!

You seem to be desperate from preventing the truth that this card runs hot from getting out there. Why? Do you own AMD stock?

No one told him the top selling 5700XT on Amazon has a 3/5 avg rating for being too hot.

No TPU graph would make me buy such a horrible product. Maybe AIB's will be better, but for now the 5700's are not great overall. Especially if you think of it as any other GPU and slap it in a case with poor airflow and/or high ambient temp.

Even third party coolers are recommended for stock Ryzen parts. AMD + 7nm = Meh.
 
No one told him the top selling 5700XT on Amazon has a 3/5 avg rating for being too hot. No TPU graph would make me buy such a horrible product. Maybe AIB's will be better, but for now the 5700's are not great overall. Especially if you think of it as any other GPU and slap it in a case with poor airflow and/or high ambient temp.
Blower model fans are actually designed for PCs with poor airflow. Mini-ITX builds and slim form factor builds use blower model fans by choice because it exhausts air outside the case.
Yes, the 5700XT cooling design is lackluster because of the blower design and low fan curve. The 5700 is more acceptable. The problem is more with the cooler design and not the technology itself because the RDNA architechture has comparable performance per watts as Nvidia's Pascal and Turing cards.
Hot temperature is not the same as overall heat output & electrical consumption. A bad cooler design is easy to overcome with AIBs aftermarket cooling. An inefficent/hot architechture with bad perf/watts and high power consumption is much harder to overcome.
So just wait for AIBs...there is no need to make a mountain out of a molehill.
Even third party coolers are recommended for stock Ryzen parts. AMD + 7nm = Meh.
Recommended by? Most of the Ryzens are 65W-95W so the stock coolers such as the Stealth and Spire are perfectly adequate. The Stealth cooler with the Ryzen 3600 maxes out out around 80'C while performing fairly quietly. The 3rd generation Ryzen 3000s can't overclock much beyond the stock boost speeds anyways so 3rd party coolers are mostly unnecessary for most of the lineup.
If you have something like the Ryzen 3600, 3600X, etc, then there is not much reason to spend more money on a third party cooler because you won't be overclocking it anyways.
 
RDNA is essentially the only Cards anybody should be buying right now, unless you are stuck on a newer G-sync monitor and need a 2080ti to push it. Otherwise... Navi.

And within a few months time, Navi will be all anyone needs because the 2080ti will become obsolete, for Gamers.


Nearly every game being made from here on out, will be made fore the Xbox Scarlet & Platstaion5 using RDNA.
 
Blower model fans are actually designed for PCs with poor airflow. Mini-ITX builds and slim form factor builds use blower model fans by choice because it exhausts air outside the case.
Yes, the 5700XT cooling design is lackluster because of the blower design and low fan curve. The 5700 is more acceptable. The problem is more with the cooler design and not the technology itself because the RDNA architechture has comparable performance per watts as Nvidia's Pascal and Turing cards.
Hot temperature is not the same as overall heat output & electrical consumption. A bad cooler design is easy to overcome with AIBs aftermarket cooling. An inefficent/hot architechture with bad perf/watts and high power consumption is much harder to overcome.
So just wait for AIBs...there is no need to make a mountain out of a molehill.

Recommended by? Most of the Ryzens are 65W-95W so the stock coolers such as the Stealth and Spire are perfectly adequate. The Stealth cooler with the Ryzen 3600 maxes out out around 80'C while performing fairly quietly. The 3rd generation Ryzen 3000s can't overclock much beyond the stock boost speeds anyways so 3rd party coolers are mostly unnecessary for most of the lineup.
If you have something like the Ryzen 3600, 3600X, etc, then there is not much reason to spend more money on a third party cooler because you won't be overclocking it anyways.

Just letting you know TPU have the review for Asus Strix 5700 XT and basically Asus use the same cooler for 2070 Super, while gaming consumption is the same for both, the result are pretty far apart:
Strix 5700XT reaches 77C at 36dBA while Strix 2070Super is only 64C at 34dBA

https://www.techpowerup.com/review/asus-radeon-rx-5700-xt-strix-oc/31.html
https://www.techpowerup.com/review/asus-geforce-rtx-2070-super-strix-oc/31.html

While 77C at 36dBA is not bad, Asus Strix model is usually the most premium one, expect worse results with cheaper model. Bigger Navi is gonna be even harder to cool.
 
Yeh, there seems to be a decent temperature difference between the Asus Strix models. The 5700XT does consume about 9W more in the gaming test, but that wattage difference in gaming is fairly small so it'd be other factors such as die size. The 5700XT does run hotter than comparable TDP Nvidia cards, but it isn't quite "hot" either as 70s'C at gaming load places it in the mid range of GPU temperatures.

While 77C at 36dBA is not bad, Asus Strix model is usually the most premium one, expect worse results with cheaper model. Bigger Navi is gonna be even harder to cool.

The cheap 2-fan 5700XT Pulse from Sapphire actually runs cooler and quieter than the ASUS STRIX according to TPU.

TPU's review of the Sapphire Pulse 5700XT version says it costs $410 (only $10 more than reference), and runs at 35 dBA at 75'C....which is quieter and cooler than the Strix at 36 dBA at 77'C. (TPU says the Sapphire model consumes a bit more electricity too)

https://www.techpowerup.com/review/sapphire-radeon-rx-5700-xt-pulse/
https://www.techpowerup.com/review/sapphire-radeon-rx-5700-xt-pulse/32.html

ASUS's 3 fan Strix doesn't seem very good if it gets outperformed by a cheap 2 fan model that barely costs more than the reference model. Or maybe they had a bad batch of Strix cards here.
 
Last edited:
Back