Counter-Strike 2 Performance Benchmark

[noel_fs]

Super interesting the performance/quality ratio of the different presets. Going from low to very high the only difference i see is AO and shadows, and you would always play with those off anyway.

 

[Chrispy_]

Interesting to see 7800xt so behind of 6800xt, and right along with rx6800. Looks like number of compute units are the bottleneck this time around, rather than clocks.

No, it's not that.

The 7600 and 6600XT are basically two versions of the same card, one with RDNA2 and one with RDNA3; The 7600 should have the edge as the game clocks are about 10% faster, as is the VRAM.

Despite that, the 7600 is underperforming by about 25% - it should be about 5-10% faster than the 6600XT and it ends up 15-20% slower.

This will be some driver optimisation where flags are set for RDNA2 optimisations but RDNA3 is running in generic/compatibility mode. I would expect a game-ready driver to address this soon, though I doubt it's a priority because the game is running ridiculously well already and the only reason to spend time on it is for the sake of purely synthetic benchmarks. 150FPS on their lowest-end RDNA3 cards at max graphics settings isn't really a "bug" that needs fixing in a hurry, especially not when turning down to competitive settings will hit the refresh rate limit of even 240Hz monitors.

Edit - and I guess it goes without saying that if you're using a $1500 monitor with a higher refresh than 240Hz, you're probably not running a low-end RX 7600!

 

[Angry]

cpu test please

 

[Enzarch]

@W1zzard What is your test pass for this? I am seeing way above these numbers for a 7900XTX @4K with matched settings and same 23.9.3 driver
Bot matches across a few maps I am seeing ~200+FPS average.
Granted, even though I ran at -10% I still have a bit higher power limit than reference, but I dont see that making a 25% performance gain

Were multiple passes run? As shader stutter on first run seems to be an issue ATM.

7800X3D & DDR5 6000 CL30
Win11 22H2

 

[ARF]

Curious to understand why the 7800XT trails the 6800XT and 6800 so significantly in this game?

Because 7800 XT is a rebranded 7700 XT, it has to trail. The question is why there are games in which RX 7800 XT is actually slightly faster than RX 6800 XT, when its specs cream the opposite.

RX 6800 XT: 4608 shaders | 288 TMUs | 128 ROPs | 128 MB L3 cache
RX 7800 XT: 3840 shaders | 240 TMUs | 96 ROPs | 64 MB L3 cache

Navi 32 is the Navi 22 successor, not the Navi 21 successor. AMD's marketing must be sacked and its members sued for misleading the whole gaming industry.

 

[Denver]

Because 7800 XT is a rebranded 7700 XT, it has to trail. The question is why there are games in which RX 7800 XT is actually slightly faster than RX 6800 XT, when its specs cream the opposite.

RX 6800 XT: 4608 shaders | 288 TMUs | 128 ROPs | 128 MB L3 cache
RX 7800 XT: 3840 shaders | 240 TMUs | 96 ROPs | 64 MB L3 cache

Navi 32 is the Navi 22 successor, not the Navi 21 successor. AMD's marketing must be sacked and its members sued for misleading the whole gaming industry.

As far as I remember, the company that created it decides what the chip will become and which segment it will fit into, not the consumer. lol

Of course AMD could just call it "RX 7800" to avoid negative comparisons.

 

[ARF]

As far as I remember, the company that created it decides what the chip will become and which segment it will fit into, not the consumer. lol

This is true but it's against their interest because the consumers can judge and can vote with their wallets - and it pretty much in all situations works against the company, not against the consumer.
I mean it's in the company best interest to present the product in its best light. When the company puts that *8** branding, the consumer naturally expects super enthusiast performance, especially when the previous generation had set the bar that far. The RX 6800 XT was truely an extraordinary offer, while RX 7800 XT is nothing but meh...

 

[stuen4y]

Thank you for the excellent review content as always W1zzard. CS:GO was notoriously a CPU test more than anything, being CPU-limited even with the fastest CPUs. Is it possible to do a CPU performance analysis on CS 2 to see how things fare? I am running an i5 11600K which is starting to show its age and I am getting between 30 and 50% GPU utilization using Present Mon with an RX 6800 even at high resolutions like 1440p. I presume a faster CPU would bring FPS (and specifically minimum FPS) up by 30-60% but I do not find any information online on the topic. As many people who play competitively can tell, having 70 or 80 FPS minimum is not enough for a fast-paced online multiplayer FPS where milliseconds decide who wins and who loses so knowing which CPU can push a certain number of minimum frames will be much appreciated. I think even a limited test with 1-2 CPUs per generation would be extremely indicative.

 

[Lewzke]

It is CPU heavy, it can spike low-mid tier 6 core CPU's to 80-90% in some moments. Probably they will fix the spikes later on.

 

[Berfs1]

I am running an i5 11600K which is starting to show its age and I am getting between 30 and 50% GPU utilization using Present Mon with an RX 6800 even at high resolutions like 1440p.

What motherboard, cooler, and RAM do you have?

 

[zo0lykas]

i tried this game back in 2021, was unplayable due mini freezes, downloaded it yesterday not much changed, still got random mini-freezes

 

[ExcuseMeWtf]

i tried this game back in 2021, was unplayable due mini freezes, downloaded it yesterday not much changed, still got random mini-freezes

How did you play CS2 in 2021?

 

[Nostras]

How did you play CS2 in 2021?

Think he meant CS:GO lol.
But GN addresses it as well in their video

Seems that Valve has some work to do, unacceptable for a high-paced shooter.

 

[ixi]

Performance is quite bad currently for CS2.

With GTX 980 TI I was able to get stable 300 fps in full hd. Now look at the bottom and middle tear for full hd... not even close to 300. Not to mention current bottom and mid tear is much stronger than 980 ti...

 

[Fraudiola01]

Because 7800 XT is a rebranded 7700 XT, it has to trail. The question is why there are games in which RX 7800 XT is actually slightly faster than RX 6800 XT, when its specs cream the opposite.

RX 6800 XT: 4608 shaders | 288 TMUs | 128 ROPs | 128 MB L3 cache
RX 7800 XT: 3840 shaders | 240 TMUs | 96 ROPs | 64 MB L3 cache

Navi 32 is the Navi 22 successor, not the Navi 21 successor. AMD's marketing must be sacked and its members sued for misleading the whole gaming industry.

and yet why does 6800XT seem to lose to the 7800XT in quite a bunch of games?

Also which do you think it’s the better option nowadays? A 6800XT or a 4070?

 

[RatusNatus]

R7 5700x, 32G 3600, RX580 MED/low config with only shadows high.
CS2 stay most of the time around 180fps , playing casual cs with 10x10 players. But theres some drops to 130 and that hinder performance hugely since the monitor is 144hz. Smokes, water and fire do have huge fps impact.
There's some stutter and it seems not related to graphics card performance. Since VOLVO hid all interpolation commands, we cant test it anymore.
Ive read somewhere that cl_showfps 1 impact fps performance 15 to 20% - so, do not use it.

AMD driver to CS2, 23.9.3 , do not support RX580 for some crazy reason since it do support RX480 so i'm not using that.

I do not get why Wizzard chose to post the minimum FPS... why's that? All cards do have drops but most of it without impact in gaming performance.

 

[wNotyarD]

I do not get why Wizzard chose to post the minimum FPS... why's that? All cards do have drops but most of it without impact in gaming performance.

May not be the case with CS2 specifically, but for some games it shows when VRAM runs out and starts being an issue.

 

[QUANTUMPHYSICS]

I hate this crap. I want CS:GO back.

 

[Chomiq]

@W1zzard what about frametimes? Have you noticed similar (up to 900 ms) spikes to what GN reported in their video?

 

[zo0lykas]

How did you play CS2 in 2021?

Download beta, like most of us

 

[Rebe1]

It would be interesting to see how different CPUs behave in this new Counter Strike

@W1zzard we put our hopes at you!

Also, any chance you checked Power consumption of GPU @Max settings? I see my 4070 Ti on max power limt hence would be cool to see how does it look like for 4090 @4K.

Download beta, like most of us

Beta was available from March 2023, in 2021 there was not even closed Alpha available for internal tests

 

[zo0lykas]

@W1zzard we put our hopes at you!

Also, any chance you checked Power consumption of GPU @Max settings? I see my 4070 Ti on max power limt hence would be cool to see how does it look like for 4090 @4K.


Beta was available from March 2023, in 2021 there was not even closed Alpha available for internal tests

 

[Remeca]

View attachment 315944

Yeah, and CS2 came out in 2012.

Perhaps those are old CSGO reviews?

 

[ixi]

Yeah, and I have counter-strike 2 since 2012. So you know. All reviews from OG went to CS2 . Including account playtime.

Yeah, and CS2 came out in 2012.

View attachment 315945

Perhaps those are old CSGO reviews?

Yep, all reviews are there since OG - csgo

View attachment 315944

Trolling is deep within ya.

 

[L'Eliminateur]

a 2023 title that does not even have DX12 or vulkan... what a joke, not touching this with a 10 foot pole.
CS was fine... back in beta 5.2 waay before the trash that became the mtx infested cs:go and because it had "up to date" graphics for when it was released.

This is a dated joke of a game, i'm not saying to add raytrace, but ANY new FPS should be as graphically heavy as the latest COD or MW, and if you have a soviet potato computer that "ran cs:go fine" tough luck