R9 Fury vs Fury Nano for Crossfire

[Cookie Monster]

Yes. How are the cards the loudest @ 0Hz? Sorry, the chart doesn't seem right.

You should disregard any measurements from say 100Hz downards. Its referred to as "pink noise" where the magnitude of this noise (i guess better wording would be its power spectral density) is inversely proportional to the frequency hence why it shoots up.

What people should be interested is the 100Hz to about 20KHz range because thats what the human ear can hear (older you are the audible frequency would be lower I guess).

If you look at that the two graphs, its easily noticeable how the Fury X actually has many high frequency peaks (especially between 15 to 20KHz which is where most of the annoying high pitched noise comes from) within the audible range compared to the EVGA hybrid. My guess is that not only is the hybrid lower in noise, its sound profile probably is more smoother if that makes sense.

 

[3DVagabond]

You should disregard any measurements from say 100Hz downards. Its referred to as "pink noise" where the magnitude of this noise (i guess better wording would be its power spectral density) is inversely proportional to the frequency hence why it shoots up.

What people should be interested is the 100Hz to about 20KHz range because thats what the human ear can hear (older you are the audible frequency would be lower I guess).

If you look at that the two graphs, its easily noticeable how the Fury X actually has many high frequency peaks (especially between 15 to 20KHz which is where most of the annoying high pitched noise comes from) within the audible range compared to the EVGA hybrid. My guess is that not only is the hybrid lower in noise, its sound profile probably is more smoother if that makes sense.

I'm aware of which frequencies are the most noticeable. Actually most peoples hearing above 15k is pretty poor, relative to frequencies below that.

Just to give some examples people might be able to relate to. The middle "C" on a piano, for example, is ~260 Hz. Middle "A", which is the halfway point on the keyboard is ~440 Hz. The highest not on a piano "C8" is ~4200 Hz. 10k+ is extremely high frequency. Also people need to realize that the difference between 10Hz and 20Hz is the same as the difference between 1KHz and 2KHz or 10KHz and 20KHz. One octave. the difference from do to do in the song do-re-mi in "The Sound of Music". So, when a scale has 1KHz segments being portrayed as equal, regardless of pitch, it's not how we hear.

 

[Headfoot]

What people should be interested is the 100Hz to about 20KHz range because thats what the human ear can hear (older you are the audible frequency would be lower I guess).

People can hear down to 20hz. Most of the thump of a 24" kick drum is at ~60hz so unless you can't hear a kick drums bass component you can hear lower than 100... you can feel frequencies in your gut even lower than 20 if its loud enough

 

[guskline]

Since you've waited this long and apparently have a spare gpu to tide you over, I would wait until the NDA for Nano is lifted to compare. Also since you want only air cooling I guess you have eliminated the Fury X since it has an AIO cooler. BTW, too bad you didn't invest in 2 EK blocks for your Titan Xs with a true Water cooling solution. I have the EK Titan X waterblock on my 980TI and it is VERY cool and quiet, even overclocking. You might want to consider 2 custom water cooled Furys.

 

[soccerballtux]

not sure why that's not a logarithmic plot but oh well

 

[Snus]

I think i would go for that card if aircool silent is the goal, reason is that 1/3 of the heatsink is open, and that in CF should give better airflow. It should theoreticly act like a "sandwich" heatsink with a fan between open heatsinks.

41db as the nano is, that is not silent, silent is always difficult, but in order to get the nano silent, u will have to change the cooler.

But if that can be more silent than Fury X, if u dont get the pumpnoise, that i dont know.

 

[3DVagabond]

I think i would go for that card if aircool silent is the goal, reason is that 1/3 of the heatsink is open, and that in CF should give better airflow. It should theoreticly act like a "sandwich" heatsink with a fan between open heatsinks.

41db as the nano is, that is not silent, silent is always difficult, but in order to get the nano silent, u will have to change the cooler.

But if that can be more silent than Fury X, if u dont get the pumpnoise, that i dont know.

The Sapphire is 2.25 slots. You need to leave space between them. I'd also add a side fan if possible.

 

[Attic]

Also worth noting that with 2xFury, like the Sapphire posted above, you can and underclock and power limit. The Fury X can be undervolted, unsure if the Fury can be.

This can have drastic reduction in fan noise for the 2 cards with subtle performance differences.

 

[alcoholbob]

Well I am going back to Nvidia. I am thoroughly unimpressed with Freesync. With Freesync on on the lower bound (45Hz range) instead of screen tearing with Vsync off I see some small ripples which is nowhere as clean as vsync. Yes, it's more responsive and might be better for a hardcore gamer but it doesn't look anything as good as vsync. I tested this in both single and crossfire configuration with the Crossover 434K and Samsung U32E850R. Also even at 45Hz with Freesync on it doesn't look anything as smooth as 60Hz. I'd say the adaptive sync claim I've heard of "40 Hz will look like 60Hz!" is a lie, or at least doesn't seem to work with the games I tested (MGSV TPP, Witcher 3, Crysis 3). Triple buffered 45Hz vsync actually looks better than 45Hz freesync IMO in terms of camera panning smoothness, although there's a slight input lag delay.

Suffice to say, I'll stick with 60Hz monitors for now and aim for the fastest cards at 4K, adaptive sync be damned. I think I'll get some quiet air cooled 980Tis.

 

[Despoiler]

Suffice to say, I'll stick with 60Hz monitors for now and aim for the fastest cards at 4K, adaptive sync be damned. I think I'll get some quiet air cooled 980Tis.

Why would you get 980tis then? The fastest cards(multiple gpu) for 4k are AMD. SLI simply cannot scale like XDMA CF. SLI also microstutters where XDMA does not.

 

[alcoholbob]

Why would you get 980tis then? The fastest cards(multiple gpu) for 4k are AMD. SLI simply cannot scale like XDMA CF. SLI also microstutters where XDMA does not.

I was getting random stutters with the Furys that I wasn't getting with the Titan Xs. I did some hardware swapping back and forth and didn't like that aspect with the DX11 games I was getting. I'm sure it'll be better come DX12 but honestly by then the new 16FF cards will be out.

 

[brandonmatic]

Well I am going back to Nvidia. I am thoroughly unimpressed with Freesync. With Freesync on on the lower bound (45Hz range) instead of screen tearing with Vsync off I see some small ripples which is nowhere as clean as vsync. Yes, it's more responsive and might be better for a hardcore gamer but it doesn't look anything as good as vsync. I tested this in both single and crossfire configuration with the Crossover 434K and Samsung U32E850R. Also even at 45Hz with Freesync on it doesn't look anything as smooth as 60Hz. I'd say the adaptive sync claim I've heard of "40 Hz will look like 60Hz!" is a lie, or at least doesn't seem to work with the games I tested (MGSV TPP, Witcher 3, Crysis 3). Triple buffered 45Hz vsync actually looks better than 45Hz freesync IMO in terms of camera panning smoothness, although there's a slight input lag delay.

Is Freesync really that unimpressive? I had convinced myself it was essential for my next monitor but maybe I should see it in person first.

 

[tential]

Is Freesync really that unimpressive? I had convinced myself it was essential for my next monitor but maybe I should see it in person first.

Depends on the monitor as well. I'm sure having a lower lower bound limit would help. If you're constantly hitting that lower bound, chances are your settings are too high. When your lower bound is that high you should be aiming for 60 fps, with dips to 50. Barely into the actual lower bound.
I'd aim for a lower bound below 40 hz if you plan on hitting 40 FPS regularly. My plan is to simply play at the EXACT settings I use now to get 60 fps, except now I won't notice the dips below 60 FPS. I wouldn't be using settings that knock me into that lower bound range with freesync it's a safety net to me, not the FPS rate you should be getting close to hitting.

Still will need to see it in person though as you said. May not be enough for me, may just need more GPU horsepower instead.

 

[alcoholbob]

Is Freesync really that unimpressive? I had convinced myself it was essential for my next monitor but maybe I should see it in person first.

I'm sure Freesync looks fantastic on a 144Hz monitor. All I'm saying anyone expecting magic or expecting 45Hz Freesync to look better than triple buffered vsync is going to be disappointed, and Freesync at the lower bounds is not going to magically look like 60Hz. I doubt adaptive sync will be relevant at 4K for a while, at least until cards can nominally drive games to 100+ fps.

 

[tential]

I'm sure Freesync looks fantastic on a 144Hz monitor. All I'm saying anyone expecting magic or expecting 45Hz Freesync to look better than triple buffered vsync is going to be disappointed, and Freesync at the lower bounds is not going to magically look like 60Hz. I doubt adaptive sync will be relevant at 4K for a while, at least until cards can nominally drive games to 100+ fps.

If the lower bound doesn't look great, or better than Vsync, then I guess I'll be looking at the fastest Nvidia card possible in my near future. Only reason I wanted freesync was for 50-55 fps to look like 60 fps. Is that possible? Because if not, then I'll be forgetting freesync monitor and just getting the fastest SLI/CF combo I can afford.

 

[3DVagabond]

If the lower bound doesn't look great, or better than Vsync, then I guess I'll be looking at the fastest Nvidia card possible in my near future. Only reason I wanted freesync was for 50-55 fps to look like 60 fps. Is that possible? Because if not, then I'll be forgetting freesync monitor and just getting the fastest SLI/CF combo I can afford.

I believe, if I understand correctly, the issue is the narrow Freesync window on his monitor. You only have 45Hz-60Hz operating range with his particular monitor.

 

[Techhog]

Well I am going back to Nvidia. I am thoroughly unimpressed with Freesync. With Freesync on on the lower bound (45Hz range) instead of screen tearing with Vsync off I see some small ripples which is nowhere as clean as vsync. Yes, it's more responsive and might be better for a hardcore gamer but it doesn't look anything as good as vsync. I tested this in both single and crossfire configuration with the Crossover 434K and Samsung U32E850R. Also even at 45Hz with Freesync on it doesn't look anything as smooth as 60Hz. I'd say the adaptive sync claim I've heard of "40 Hz will look like 60Hz!" is a lie, or at least doesn't seem to work with the games I tested (MGSV TPP, Witcher 3, Crysis 3). Triple buffered 45Hz vsync actually looks better than 45Hz freesync IMO in terms of camera panning smoothness, although there's a slight input lag delay.

Suffice to say, I'll stick with 60Hz monitors for now and aim for the fastest cards at 4K, adaptive sync be damned. I think I'll get some quiet air cooled 980Tis.

You know that AMD recommends having v-sync on with freesync, right?

 

[Mr Evil]

...I tested this in both single and crossfire configuration with the Crossover 434K and Samsung U32E850R. Also even at 45Hz with Freesync on it doesn't look anything as smooth as 60Hz...

I also have a U32E850R, and that doesn't match my experience. I did a lot of testing in Unreal Tournament 3, as it's where I notice lag most easily, and it has a built-in frame rate limiter. I compared 45 fps with FreeSync against 60 fps without, and 45 fps felt better. Only when it drops below 40 fps does it go bad, with obvious stuttering. I also tested against 85Hz without FreeSync (overclocked with "Custom Resolution Utility", although I had to lower the resolution to 2560x1440), and that started to feel only slightly better than FreeSync at 45 fps.

The 40-60Hz range is admittedly small, but it can be improved by driver modding; I have mine at 33-60Hz, which is a massive improvement. In, for instance The Witcher 3, which is hard to keep >40Hz at 4k with decent settings, it doesn't look quite as good as fixed 60Hz, but the responsiveness is still amazing.

This is with a plain Fury, and I kept V-sync on in all the tests.

 

[tential]

I also have a U32E850R, and that doesn't match my experience. I did a lot of testing in Unreal Tournament 3, as it's where I notice lag most easily, and it has a built-in frame rate limiter. I compared 45 fps with FreeSync against 60 fps without, and 45 fps felt better. Only when it drops below 40 fps does it go bad, with obvious stuttering. I also tested against 85Hz without FreeSync (overclocked with "Custom Resolution Utility", although I had to lower the resolution to 2560x1440), and that started to feel only slightly better than FreeSync at 45 fps.

The 40-60Hz range is admittedly small, but it can be improved by driver modding; I have mine at 33-60Hz, which is a massive improvement. In, for instance The Witcher 3, which is hard to keep >40Hz at 4k with decent settings, it doesn't look quite as good as fixed 60Hz, but the responsiveness is still amazing.

This is with a plain Fury, and I kept V-sync on in all the tests.

ok that's interesting. I basically want 4K freesync so I can play 4K at 60hz, but have dips below that and be ok. I expect dips to 50 FPS, but rarely 45 fps. My plan is Crossfire R9 290s, then crossfire arctic islands since I'll definitely need it as I know I'll want to crank up settings.

Have money set aside for the Wasabi UHD650 but I'm starting to worry that 1080p 120hz DSR might be better for the high refresh rate. Decisions decisions! I wish there was a place I could actually see this in action.

 

[Techhog]

I also have a U32E850R, and that doesn't match my experience. I did a lot of testing in Unreal Tournament 3, as it's where I notice lag most easily, and it has a built-in frame rate limiter. I compared 45 fps with FreeSync against 60 fps without, and 45 fps felt better. Only when it drops below 40 fps does it go bad, with obvious stuttering. I also tested against 85Hz without FreeSync (overclocked with "Custom Resolution Utility", although I had to lower the resolution to 2560x1440), and that started to feel only slightly better than FreeSync at 45 fps.

The 40-60Hz range is admittedly small, but it can be improved by driver modding; I have mine at 33-60Hz, which is a massive improvement. In, for instance The Witcher 3, which is hard to keep >40Hz at 4k with decent settings, it doesn't look quite as good as fixed 60Hz, but the responsiveness is still amazing.

This is with a plain Fury, and I kept V-sync on in all the tests.

And that's why you had better results.

 

[brandonmatic]

And that's why you had better results.

Isn't the issue with Vsync that if you drop out of the Freesync range into Vsync you get input lag and choppiness? At that point I think I'd prefer tearing to be honest.