Horizon Zero Dawn [PS4, PC]

Why would you need the same gameplay paths when every gameplay path on PS4 is the same 30 fps? You get rare 1-2 fps drops with an average of like 29.998. Thats worlds apart from the 22 fps avg a 7850 gets at the same settings.

I'd rather rely on benchmarks then individual gameplay moments, in special for a game that's in a bad state at the moment. It would be a rather rare sight to see a hd7850/70 performing much worse then the base PS4. My 7870 outperforms the base ps4 at the same settings as the Pro aside from resolution, a 7850, as the benchmark shows, is ballpark-on-par.

Offcourse people and benchmarks could be lying and wrong, but i won't go there in the technical thread.
 
I'd rather rely on benchmarks then individual gameplay moments, in special for a game that's in a bad state at the moment. It would be a rather rare sight to see a hd7850/70 performing much worse then the base PS4. My 7870 outperforms the base ps4 at the same settings as the Pro aside from resolution, a 7850, as the benchmark shows, is ballpark-on-par.

Offcourse people and benchmarks could be lying and wrong, but i won't go there in the technical thread.
Wow. How is this real life.
 
I find the fact that a 7850 is getting anywhere close to the PS4's performance pretty remarkable. Whether its roughly as fast as shown by the benchmark, or a bit slower as shown in the video Techuse posted (both great pieces of information btw) it's still a remarkable feat.

Remember that a 7850 is a slower GPU than what's in the PS4 with less memory bandwidth, a quarter the VRAM, and an older IP. Its also long since left AMD's priority list for driver optimisations.

That, coupled with the lower level coding on a console, especially in a game like this, makes it pretty stunning IMO that the 7850 is anywhere near the base PS4.
 
I find the fact that a 7850 is getting anywhere close to the PS4's performance pretty remarkable. Whether its roughly as fast as shown by the benchmark, or a bit slower as shown in the video Techuse posted (both great pieces of information btw) it's still a remarkable feat.

Remember that a 7850 is a slower GPU than what's in the PS4 with less memory bandwidth, a quarter the VRAM, and an older IP. Its also long since left AMD's priority list for driver optimisations.

That, coupled with the lower level coding on a console, especially in a game like this, makes it pretty stunning IMO that the 7850 is anywhere near the base PS4.

Yes exactly, also, AMD probably has yet to, if any, optimized drivers for that type of 2012 hardware. A 7870 probably is a better match for the PS4, in pure TF performance atleast.
My 7870 performs better in almost any game tested so far, even doom eternal and the pre- sony exclusives.
 
I find the fact that a 7850 is getting anywhere close to the PS4's performance pretty remarkable. Whether its roughly as fast as shown by the benchmark, or a bit slower as shown in the video Techuse posted (both great pieces of information btw) it's still a remarkable feat.

Remember that a 7850 is a slower GPU than what's in the PS4 with less memory bandwidth, a quarter the VRAM, and an older IP. Its also long since left AMD's priority list for driver optimisations.

That, coupled with the lower level coding on a console, especially in a game like this, makes it pretty stunning IMO that the 7850 is anywhere near the base PS4.

I think it would be useful to put the specs on the table for this to make it easier to compare.

HD7850

1.76 TFLOPS
153.6GB/s Bandwidth
2GB VRAM

PS4

1.84 TFLOPS
176.00GB/s Bandwidth
somewhere between 0 and 8GB of VRAM

So the detla between the TFLOPS of the two is ~80GFLOPS
Id say the RAM bandwidth is pretty much a wash because the PS4s is shared

I couldnt find concrete numbers on Horizons RAM usage or the current reservation on the PS4 so its hard to talk to that.

But based on the specs above I wouldnt say that the HD7850 running a game at 720P that the PS4 runs at 30ish 1080P is 'anywhere close'.

I also feel its pretty disengious to compare the benchmark results of a video card against the PS4 when we don't know what the PS4 got in the same benchmark, the only real comparasion can be made to gameplay.

Yes exactly, also, AMD probably has yet to, if any, optimized drivers for that type of 2012 hardware. A 7870 probably is a better match for the PS4, in pure TF performance atleast.
My 7870 performs better in almost any game tested so far, even doom eternal and the pre- sony exclusives.

a HD 7870 is no where near closer to the PS4 in pure TFLOPS

PS4
1.84 TFLOPS

HD7850
1.76 TFLOPS

HD7870
2.56 TFLOPS

The delta between the PS4 and HD7870 is nearly 10x the difference between the PS4 and the HD7850, so how is it a better comparasion to the PS4?.
 
And yet, an 7850 is the closest match to the base ps4 in just about any game. We use benchmarks since they are there for a reason. In special with hzd port since there are problems GG is currently working on.
Aside from that, for some there are not the performance problems, for others there are.

A 7850 matching the base ps4 @ original/1080p settings is quite impressive, but at the same time right where it should be.
 
VRAM usage for PS4 will be less than 5Gig as it needs to use space for program code (game usage is 5Gig total). The rest of the ram is used by the OS/Dashboard.
 
VRAM usage for PS4 will be less than 5Gig as it needs to use space for program code (game usage is 5Gig total). The rest of the ram is used by the OS/Dashboard.

So closer to 2x than 4x the VRAM. I also suspect that the 2GB is more then okay for the HD7850 unless AMD for some reason purposely bandwidth / VRAM limited there video cards because the HD7870 has the same VRAM configuration.

And yet, an 7850 is the closest match to the base ps4 in just about any game. We use benchmarks since they are there for a reason. In special with hzd port since there are problems GG is currently working on.
Aside from that, for some there are not the performance problems, for others there are.

A 7850 matching the base ps4 @ original/1080p settings is quite impressive, but at the same time right where it should be.

But it doesn't according to the in game video?. The benchmark shows it at 30FPS @ 1080P, and we have no idea how the PS4 would perform in the benchmark.

In game requires the resolution to be dropped too 720P to be a stable frame rate, compared to the PS4s mostly stable 1080P.
 
I think it would be useful to put the specs on the table for this to make it easier to compare.

HD7850

1.76 TFLOPS
153.6GB/s Bandwidth
2GB VRAM

PS4

1.84 TFLOPS
176.00GB/s Bandwidth
somewhere between 0 and 8GB of VRAM

So the detla between the TFLOPS of the two is ~80GFLOPS
Id say the RAM bandwidth is pretty much a wash because the PS4s is shared

I couldnt find concrete numbers on Horizons RAM usage or the current reservation on the PS4 so its hard to talk to that.

But based on the specs above I wouldnt say that the HD7850 running a game at 720P that the PS4 runs at 30ish 1080P is 'anywhere close'.

You forgot to mention the lower IP level. There's also numerous examples in modern games of the same GPU with more VRAM achieving significantly higher performance than it's lower VRAM model. So that could certainly be having a significant impact on the performance. As I mentioned above, given this GPU is likely not receiving any serious driver optimizations these days, and isn't even officially supported by the game, and thus unlikely to have gotten any developer love, I do think this is a very impressive showing.

I also feel its pretty disengious to compare the benchmark results of a video card against the PS4 when we don't know what the PS4 got in the same benchmark, the only real comparasion can be made to gameplay.

It's still useful to consider as an additional data point. We currently have 2 data points. A video of the repeatable and well understood built in benchmark, and commentary stating that a mix of low and medium settings at 720p gives a frame rate of 30+fps with AMD's Fidelity FX and 16x AF set in the drivers. Neither of those are precise comparison points for the PS4 but both seem worthy of inclusion in any analysis.
 
You forgot to mention the lower IP level. There's also numerous examples in modern games of the same GPU with more VRAM achieving significantly higher performance than it's lower VRAM model. So that could certainly be having a significant impact on the performance. As I mentioned above, given this GPU is likely not receiving any serious driver optimizations these days, and isn't even officially supported by the game, and thus unlikely to have gotten any developer love, I do think this is a very impressive showing.

I thought the IP level was roughly southern islands? that being the HD7850 level?.

It's still useful to consider as an additional data point. We currently have 2 data points. A video of the repeatable and well understood built in benchmark, and commentary stating that a mix of low and medium settings at 720p gives a frame rate of 30+fps with AMD's Fidelity FX and 16x AF set in the drivers. Neither of those are precise comparison points for the PS4 but both seem worthy of inclusion in any analysis.

Its a useful data point but its not a useful comparison to the PS4 because there is nothing to compare it to.
 
Back
Top