Yes you need.
I've tried the game during the server stress test, and at the time, the game didn't start on my Phenom II CPU, crashing to desktop, while on my laptop with an intel CPU with SSE 4.2 it did start.
This might get fixed on a patch or final version (happened before with Mirrors Edge Catalyst which the beta was lacking support and the final version did work), but seeing how AC Origins never received a patch, I do think that ubisoft will simply abandon non-SSE 4.1 CPUs which is weird because Dunia engine it's an old engine.
On the other side Final Fantasy XV just added support to Phenom II CPUs in a patch, which was also another game that lacked support of non-SSE 4.1 CPUs.
Cheers
Thanks. I got a response back and was told because of DRM! Will wait a month or two and no support I will build a new rig. It's disgusting that people have to buy a new CPU just because it can't run DRM. The Ryzen I build better give me 10 years of use like my old build lol.
From Guru3D's own testing:
At 1080p Ultra, The VEGA 56 is 3% slower than the GTX 1070 Ti, while the VEGA 64 is 3% slower than the GTX 1080.
At 1140p Ultra, The VEGA 56 is 1% faster than the GTX 1070 Ti, while the VEGA 64 is 7% faster than the GTX 1080.
At 2160p Ultra, The VEGA 56 is 2% faster than the GTX 1070 Ti, while the VEGA 64 is 9% faster than the GTX 1080.
Typical stupidity of a Fanboy is the failure to recognize the validity of a legit question asked by an enthusiast, which by no means is an attempt to trash their beloved manufacturer. AMD focused a lot on Rapid Packed Math, Primitive Shaders and what not during VEGA's launch. So it is a legit question to understand and quantify the performance benefit/uplift from such features, when implemented by a game developer. As from Guru3D's own figures, the VEGA 64 numbers are somewhat impressive, but the VEGA 56 numbers aren't all that better.
sykozis:
If the framerates are achieved without a loss of visual quality, then what does it really matter how the framerates are achieved?
I'll quote a portion of Hilbert's own article:
[Spoiler]
AMD effects
In our overview, we talked a bit about AMD having the ability (with Vega) to have several render targets in a scene to render at lower precision (fp16 versus fp32). AMD calls this feature fp16 Rapid Packed Math (floating point 16). According to AMD this will increase performance. Rapid Packed Math basically halves the floating point calculations for the data request, resulting in a faster turnaround time of that request/data, however with less precision and thus quality in some form. Basically, half-precision would be applied in segments where it really isn't needed. AMD has not revealed anything specific as to what and where exactly the feature can be used. Far Cry 5 supports this, specifically with Vega and newer future GPUs, and as you have been able to see, the Vega cards perform well
[/Spoiler]
Do you see the part in bold? There is no such thing as free performance, the performance always comes at a cost or a workaround.
Prince Valiant:
Perhaps they're too used to losing performance with Gameworks that getting a free improvement must be bad 😀?
I must've stirred up the Fanboy's nest. If you go through the history of my posts, you'll see that I have been using hardware from both nvidia and AMD, and I am all for free improvements/performance. I have been using AMD GPUs for perhaps longer than most have, not sure why you've this silly attitude.
With the stupidity of fanboism aside, a benchmark is a means to demonstrate the capabilities of hardware from different manufacturers, when they are compared apples to apples. If this had been a Gameworks title, surely for the sake of a fair comparison, the gameworks features would've been disabled. For the idea of a fair comparison from my understanding, I asked a simple question; a simple inquiry on how much does VEGA benefit from Rapid Packed Math, to learn the extent of it. It appears that asking for a fair comparison or having the simplest inquiry regarding the extent of the difference a (overhyped) feature makes can irk some people, regardless whether they are Green or Red. Fanboism is the highest form of stupidity when it comes to technology.
Typical stupidity of a Fanboy is the failure to recognize the validity of a legit question asked by an enthusiast, which by no means is an attempt to trash their beloved manufacturer.
A fanboy you say?
NVidia cards I've owned: GeForce MX200, GeForce MX400, some Riva TNT2 based card, GeForce 5600XT, 5700LE, 6200, 6800, 7300, 7600GT (x2), 8600GT (x2), 9600GT, GTS250, GT220, GTS240, GTX275, GTX460 (x2), GTX560, GTX660 (x2), GTX970
AMD/ATI cards I've owned: 3D Rage Pro, 9200SE (x2), 9600XT, x700 Pro, HD2400, HD4850, HD7870, HD7950, R5 240, RX470.....
I think if I were going to be a "fanboy"....based on my purchasing history....I'd be leaning more towards NVidia than AMD. I mean, 23 NVidia cards vs 11 AMD/ATI cards....
Well, this: "however with less precision and thus quality in some form." concerns water right?
Can any difference be detected?
If not then doing tricks like this should be encouraged, we do not need Crysis 2 tesselation on places where it is not even seen just to sell more graphiccards for changes that are impossible to detect.
If you can't see the difference during normal game play, without having to search for the differences or take screenshots to compare, why worry about the difference? I thought the point of videos games was to play and enjoy the game, not stare at the scenery....
If you can't see the difference during normal game play, without having to search for the differences or take screenshots to compare, why worry about the difference? I thought the point of videos games was to play and enjoy the game, not stare at the scenery....
I've read about this feature before and always interpreted it as being meant for replacing FP32 in areas where they can do so without impacting the quality
The reason we even see this is probably because AMD and Ubisoft brought it over from the PS4 Pro version.
I am here to offer a different point of view, you see gameworks is a set of features, like any other features in a game, it can be enabled/disabled. But RPM (Rapid Packed Math) or shader intrinsic functions are not your typical features rather it's an optimization. The kinda optimization that can be seen in consoles. If consoles (specially ps4 pro) can take advantage of such functions without needing any gameworks features and still deliver gorgeous looking titles like Horizon Zero Dawn, Uncharted 4, God of War (unreleased) then imho its totally worth it to use such optimizations.
I mean just look at the Far Cry 5 tech analysis (for consoles) by digital foundry https://youtu.be/xuKpA2iORd4 they get away without relying on heavy tessellation yet created realistic looking stone walls, ground, dirt, terrain, tree bark, trunk etc. Ain't that impressive in a world where you need to slap on "gameworks" to make a game look good and realistic?
You do know the Crysis 2 tesselation thing was proven to be a false narrative? The mesh is there but if it’s not on the screen it’s not rendered.
You mean the water? But there are other things which are extremely highpoly:
https://techreport.com/review/21404/crysis-2-tessellation-too-much-of-a-good-thing/2
It's like when Crytek claimed you needed a 64bit CPU and ofc OS to play Far Cry 1 with all the bells and whistles but somebody quickly made it work with 32bit. Or DX10 features which could be enabled in DX9 with just a few ini tweaks.
These sort of things are of course done to sell hardware.
In our overview, we talked a bit about AMD having the ability (with Vega) to have several render targets in a scene to render at lower precision (fp16 versus fp32). AMD calls this feature fp16 Rapid Packed Math (floating point 16). According to AMD this will increase performance. Rapid Packed Math basically halves the floating point calculations for the data request, resulting in a faster turnaround time of that request/data, however with less precision and thus quality in some form. Basically, half-precision would be applied in segments where it really isn't needed. AMD has not revealed anything specific as to what and where exactly the feature can be used. Far Cry 5 supports this, specifically with Vega and newer future GPUs, and as you have been able to see, the Vega cards perform well
Do you see the part in bold? There is no such thing as free performance, the performance always comes at a cost or a workaround.
Depends on what they are using Rapid Packed Math for.
If they are using it properly for calculations where FP32 is overkill, and FP16 gives enough precision, then there is no image degradation, just a potential speedup of having two ops executed (in FP16 precision) instead of just one ops(in FP32 precision)
OTOH if they are using it indiscriminately or hastily, then it's no different then FP16 demotion.
There is a good write-up on RPM on anand:
Arguably AMD’s marquee feature from a compute standpoint for Vega is Rapid Packed Math. Which is AMD’s name for packing two FP16 operations inside of a single FP32 operation in a vec2 style. This is similar to what NVIDIA has done with their high-end Pascal GP100 GPU (and Tegra X1 SoC), which allows for potentially massive improvements in FP16 throughput.
The purpose of integrating fast FP16 and INT16 math is all about power efficiency. Processing data at a higher precision than is necessary unnecessarily burns power, as the extra work required for the increased precision accomplishes nothing of value. In this respect fast FP16 math is another step in GPU designs becoming increasingly min-maxed; the ceiling for GPU performance is power consumption, so the more energy efficient a GPU can be, the more performant it can be
Taking advantage of this feature, in turn, requires several things. It requires API support and it requires compiler support, but above all it requires code that explicitly asks for FP16 data types. The reason why that matters is two-fold: virtually no existing programs use FP16s, and not everything that is FP32 is suitable for FP16. In the compute world especially, precisions are picked for a reason, and compute users can be quite fussy on the matter. Which is why fast FP64-capable GPUs are a whole market unto themselves. That said, there are whole categories of compute tasks where the high precision isn’t necessary; deep learning is the poster child right now, and for Vega Instinct AMD is practically banking on it.
As for gaming, the situation is more complex still. While FP16 operations can be used for games (and in fact are somewhat common in the mobile space), in the PC space they are virtually never used. When PC GPUs made the jump to unified shaders in 2006/2007, the decision was made to do everything at FP32 since that’s what vertex shaders typically required to begin with, and it’s only recently that anyone has bothered to look back. So while there is some long-term potential here for Vega’s fast FP16 math to become relevant for gaming, at the moment it doesn’t do much outside of a couple of benchmarks and some AMD developer relations enhanced software. Vega will, for the present, live and die in the gaming space primarily based on its FP32 performance.
Though overall it’s important to keep in mind here that even in the best case scenario, only some operations in a game are suitable for FP16. So while FP16 execution is twice as fast as FP32 execution on paper specifically for a compute task, the percentage of such calculations in a game will be lower.
New water engine in Far Cry 5: FP16 calculations on AMD graphics cards http://www.pcgameshardware.de/screenshots/original/2018/03/21-Angel-Bug-pcgh_b2article_artwork.JPG Source: PC Games Hardware
21.03.2018 at 17:45 Coming up next week is Far Cry 5, Ubisoft's newest open-world adventure. At GDC 2018 in San Francisco, the developers introduced a completely new water engine that they developed for the game. And finally, you also know what effects on FP16 calculations fall back and thus AMD graphics card especially.
Water plays a big role in every far cry. The first action adventure developed by Crytek took us to a wet tropical island. Also in the successors it went again and again into the water, which was constantly improved by the programmers.
But for Far Cry 5 , the developers wanted to dig deeper into the bag of tricks. "Montana, the scene of our game, has plenty of rivers and rapids - which we would not have been able to do so elegantly with our old shaders," says Branislav Grujic, 3D Team Lead Programmer at Ubisoft Toronto. "Add to that more than 50 waterfalls that we used to make by hand, which would have taken enormous time, and sulfur mines that would not have been possible with the old routines."
A game world from 4,096 sectorshttp://www.pcgameshardware.de/screenshots/430x/2018/03/23-Christian-Cutocheras-und-Branislav-Grujic-pcgh.JPG 23 Christian Cutocheras and Branislav Grujic Source: PC Games Hardware The world of Far Cry 5 is divided into sectors of which there are 64 times 64, ie 4,096 pieces. One of many lakes consumes about 20 to 30 of them. And because there are so many waters in the state of Montana, that would have been a lot of hard work by hand.
The developers therefore resorted to procedural irrigation: For example, boundary points are noted on the terrain for a river, then automatically filled with water using the flow tool. Procedural waterfalls rush into the depths of foam thanks to Foam Mapping, while a helicopter over a sulfur pit releases particles from it and whirls them up into the air.
Render path and whitecaps
The render path, which is responsible for the realistic representation of the water, uses some tricks. Clever occlusion calculations and the subdivision into near and far rendering do not require the calculation of every drop of water in the highest level of detail. In the previously hand-drawn normal maps of the waves on the water surface set the team in Far Cry 5 on their automatic generation by noise function, which is based on the Brownian molecular motion.
On rocks and coastal areas often form foam crowns, which are automatically created again with noise features. Here, the developers resorted to a trick: The Foam Maps of the entire game world are "baked" in the game world, because their individual calculation would have deducted too much computing power from other processes and not least the actual game. Flow maps are also created automatically, but calculated dynamically. They are based on spline and flood-fill routines, which are displayed in greater detail near the player than at a greater distance.
Faster thanks to half accuracy
Bug cleaning water smoothing, lighting and refraction also require a lot of computing power. Christian Cutocheras, Member of Technical Staff, AMD, who helped Ubisoft designers with advice and support, explains a trick the programmers used: "To optimize the shaders, we're reducing the computational accuracy to 16-bit Half Precision." However, that was not allowed to happen automatically: developers and the QA team had to scour the entire game world for areas in which the reduced accuracy did not strike the players.
Remember, Radeon RX Vega graphics cards are particularly interesting for half-precision FP16 calculations, as Vega 10 can run them at twice the speed ("Rapid Packed Math"). Polaris can also run FP16 natively, but not with double FP32 throughput, thus saving only resources.
For further puzzle tasks created differently deep, colliding waters, whose occasional edges were made somewhat blurred by Compute Shader, to ensure a smoother transition. And then there was the fishing line, which can be seen both in the front and in the back of the picture and happily danced up and down the tail ...
Two milliseconds per frame
The goal of the developers was to complete the water calculations in two milliseconds per frame. At 1.919 milliseconds, they reached their destination, and by using Async Compute, they were able to reduce that time by 0.582 to 1.337 milliseconds. However, the "most expensive" operations, the tessellation and the joining of the overall picture, could not be impressed.
Impressive, however, that the water system is only one part of the big picture - in the finished game, for example, there are particles and the actions of players and NPCs, who in turn put their stamp on the environment.
My RX 580 score https://imgur.com/download/FelCnvs
minimum: 66
average: 73
max: 86
frames: 4231
RX 580 aorus 1380Mhz/2000Mhz stock
just 1 frame less than the Core i7-5960X Extreme at 4.2Ghz
My RX 580 score https://imgur.com/download/FelCnvs
minimum: 66
average: 73
max: 86
frames: 4231
RX 580 aorus 1380Mhz/2000Mhz stock
just 1 frame less than the Core i7-5960X Extreme at 4.2Ghz
You found that you're rx580 is epic bottleneck in FC5. The cpu is waiting for the slow gpu 😛
60fps in 2018 is cancer to the eyes 😛
I "need" 100+ fps for my monitor
As with all things monitor related, it's entirely subjective. If you are playing at 100+ hz, then you aren't playing at 4k - i prefer playing at 4k 60hz to playing at higher hz at lower res 🙂
When latest console is doing just fine at 4k but your $600 card starts to get into the territory of integrated graphics frame rates, you know something is fucked up.
Well even the xbox x is running the game upscaled from 1440p, not native 4k, and it does so at 30 fps 🙂
I disabled Hyper Threading (HT) on my i7-7700K and mouse responsives has been increased (less input- or mouse-lag). https://static5.cdn.ubi.com/u/ubiforums/20130918.419/images/smilies/cool.png
Also FPS is higher now, earlier 3440x1440 Ultra i got 74/87/100 FPS on FC5 bench, with 99% GPU usage.
Now i got 78/90/105 FPS, at same 99% GPU usage.
CPU now has sometimes all 4 cores 100% usage with gaming, but it still works just flawlessly, no lag or what-so-ever issues.
i7-7700K @5400
1080Ti @2000/11800
64Gb @3600 15-15-15
[Spoiler]
AMD effects
In our overview, we talked a bit about AMD having the ability (with Vega) to have several render targets in a scene to render at lower precision (fp16 versus fp32). AMD calls this feature fp16 Rapid Packed Math (floating point 16). According to AMD this will increase performance. Rapid Packed Math basically halves the floating point calculations for the data request, resulting in a faster turnaround time of that request/data, however with less precision and thus quality in some form. Basically, half-precision would be applied in segments where it really isn't needed. AMD has not revealed anything specific as to what and where exactly the feature can be used. Far Cry 5 supports this, specifically with Vega and newer future GPUs, and as you have been able to see, the Vega cards perform well
[/Spoiler] Do you see the part in bold? There is no such thing as free performance, the performance always comes at a cost or a workaround. I must've stirred up the Fanboy's nest. If you go through the history of my posts, you'll see that I have been using hardware from both nvidia and AMD, and I am all for free improvements/performance. I have been using AMD GPUs for perhaps longer than most have, not sure why you've this silly attitude. With the stupidity of fanboism aside, a benchmark is a means to demonstrate the capabilities of hardware from different manufacturers, when they are compared apples to apples. If this had been a Gameworks title, surely for the sake of a fair comparison, the gameworks features would've been disabled. For the idea of a fair comparison from my understanding, I asked a simple question; a simple inquiry on how much does VEGA benefit from Rapid Packed Math, to learn the extent of it. It appears that asking for a fair comparison or having the simplest inquiry regarding the extent of the difference a (overhyped) feature makes can irk some people, regardless whether they are Green or Red. Fanboism is the highest form of stupidity when it comes to technology.