Hitman 2 PC Performance Review

[WYP]

A high-quality port, or a PC game that is in dire need of extra polish?



Read more about Hitman 2's PC Performance.

 

[AlienALX]

hahaha that pic is hilarious you've got Intel shooting AMD, then Nvidia throwing AMD over the banister. hahahaha !!

 

[WYP]

hahaha that pic is hilarious you've got Intel shooting AMD, then Nvidia throwing AMD over the banister. hahahaha !!

It's good to have fun with the images.



Apologies for how late this one is peeps, I was unwell last week so things got a little delayed.

 

[BFBC2033]

Your testresults are quiet unusual, just take a look around. For example Setve from hw undboxed got his vega 56 infront of a 1070ti, yes thats a bit high but way more in line than what you got. I would strongly suggest you to look at your testsetup. Your test is the only one with such a massive deficite for AMD. Also those childish pictures leave a untrustworthy tast behind.

 

[WYP]

Your testresults are quiet unusual, just take a look around. For example Setve from hw undboxed got his vega 56 infront of a 1070ti, yes thats a bit high but way more in line than what you got. I would strongly suggest you to look at your testsetup. Your test is the only one with such a massive deficite for AMD. Also those childish pictures leave a untrustworthy tast behind.

I will have a look the HW Unboxed's findings and see how his test location compares to the one that we used.

We use a location that is much later in the game and represents some of the worst performance that the game can offer. I will look at the game again to see if anything changes, but each graphics card has a clean driver install, so it isn't that.

The mission HW unboxed used appears to be one of the lesser demanding areas, but I will report back when I look into things a little more.

 

[NeverBackDown]

Your testresults are quiet unusual, just take a look around. For example Setve from hw undboxed got his vega 56 infront of a 1070ti, yes thats a bit high but way more in line than what you got. I would strongly suggest you to look at your testsetup. Your test is the only one with such a massive deficite for AMD. Also those childish pictures leave a untrustworthy tast behind.

It all depends on the mission. Reviews have never been apples to apples even with the same hardware. Not sure why you expect anything else?

As for the pictures they are so irrelevant that to assume it makes an article untrustworthy is to me quite childish that it bothers someone. I only mention this because you come off as an AMD guy who is upset the vega 56 and the cover article photo both show AMD falling behind. I mean we already know a Vega 56 is slower than a 1070ti, so I again don't know why you expect anything else mate?

Food for thought, maybe Steve's results are a bit off and not WYPs? I would be more suspect a 1070ti lost a Vega 56 than anything.
Not attacking you. Just curious as to why you think the review was bad when it honestly is one of the very few performance reviews on the internet that I found. I looked around as you suggested and found just this article. Nothing else comparing really anything(outside of consoles from Digital Foundry).

 

[kardeon]

Appropriate scene ?

First of all, I always do appreciate your reviews and analyzes. Your are my favorite tech website so far.
But if i regret that you have choosen one of the most demanding scene to benchmark hitman 2. At the end of the day this benchmark doesn't reflect the real experience that we get, playing the all game.
I don't have reach yet the scene of your benchmark. But since the beginning i play easely 1440p ultra 60 fps vsync ON with a vega 64. I precise that i really don't care who is the best between Nvidia or Amd and it is not the point of my comment.

With all the respect i have for your work, i think the Hardware Unboxed's hitman 2 benchmark, seems to be more revelant for the final consumers whatever the gpu they own.

A benchmark must show the average gaming experience throughout the game. In my humble opinion, the scene you have choosen is not the most appropriate for that.
Thanks for attention and your hard work.

 

[ImprovizoR]

Games like Hitman where each level is its own little world are quite rare these days. Perhaps each level should be benchmarked separately.


But if you're going to benchmark only one level, then I think it should be the one that requires the most amount of horsepower. That will give people the worst case scenario, which is more useful IMHO. I remember when The Witcher 3 was released and most benchmarks focused on the starting area, which didn't even seem to require more than 4gb of RAM. Things change drastically once you leave White Orchard and reach Velen. Then they change even further once you reach Novigrad.

 

[WYP]

First of all, I always do appreciate your reviews and analyzes. Your are my favorite tech website so far.
But if i regret that you have choosen one of the most demanding scene to benchmark hitman 2. At the end of the day this benchmark doesn't reflect the real experience that we get, playing the all game.
I don't have reach yet the scene of your benchmark. But since the beginning i play easely 1440p ultra 60 fps vsync ON with a vega 64. I precise that i really don't care who is the best between Nvidia or Amd and it is not the point of my comment.

With all the respect i have for your work, i think the Hardware Unboxed's hitman 2 benchmark, seems to be more revelant for the final consumers whatever the gpu they own.

A benchmark must show the average gaming experience throughout the game. In my humble opinion, the scene you have choosen is not the most appropriate for that.
Thanks for attention and your hard work.

Games like Hitman where each level is its own little world are quite rare these days. Perhaps each level should be benchmarked separately.


But if you're going to benchmark only one level, then I think it should be the one that requires the most amount of horsepower. That will give people the worst case scenario, which is more useful IMHO. I remember when The Witcher 3 was released and most benchmarks focused on the starting area, which didn't even seem to require more than 4gb of RAM. Things change drastically once you leave White Orchard and reach Velen. Then they change even further once you reach Novigrad.

This is a big problem amongst PC performance analysis pieces, as many of them are based on the performance of a game's opening area, 1st or second mission and are often in an area where things are not as complex or demanding as late game areas. For Battlefield V I played through every war story before testing, just to make sure I selected a good testing area that represented general performance and locations with performance dips.

A prime example of this is Assassin's Creed: Origins or Odyssey, as CPU loads will hit their peak in large towns on cityscapes whereas other locations are more representative of general gameplay. In Shadow of the Tomb Raider, we also see much higher demands on CPUs than in the game's benchmark, which is why gameplay is always preferable for benchmarking. Scenarios matter.

When I do my work I play the game for a few hours to get a feel for how it works. With Hitman 2 the issue is that each level is almost a came within itself, so selecting a location can often be tricky. My rule of thumb on the GPU side is to go for an area that is firstly repeatable and has limited variation and represents the game in a location that is reasonably representative of high loads. Note that this is often not the worst that the game has to offer, but acts as a "it if stays over 60 here, it will stay over 60 pretty much anywhere".

As a developer, IO Interactive appears to take criticisms well, and release a lot of patches and content updates for their recent games. Patch 2.11 already addresses our complaints about CPU-related settings options, which means that I have some work to do retesting this game and updating the CPU aspects of this review.

 

[MacLeod]

I think it's proper to test GPU's on a games most demanding sections. There are sections in Crysis 3 that my old 290x could get 80-100 fps at max settings... Then there were entire sections where it would be in the 40's.

I would rather a review show me the worst case scenario. If a review shows me I can get 60 fps with a 1070 at 1440, if it's tested in the most demanding sections then I know that I can expect 60 fps anywhere in that game. If the review is based on an easy section then I'm gonna find myself pausing the game in order to change settings cause I'm crawling thru at 30 fps in the most demanding parts of the game.

 

[AlienALX]

I will take worst case scenario reviews over buttered up *any* day thanks. That is why I am still bloody here !

 

[Warchild]

I think it's proper to test GPU's on a games most demanding sections. There are sections in Crysis 3 that my old 290x could get 80-100 fps at max settings... Then there were entire sections where it would be in the 40's.

I would rather a review show me the worst case scenario. If a review shows me I can get 60 fps with a 1070 at 1440, if it's tested in the most demanding sections then I know that I can expect 60 fps anywhere in that game. If the review is based on an easy section then I'm gonna find myself pausing the game in order to change settings cause I'm crawling thru at 30 fps in the most demanding parts of the game.

Actually I prefer a balance. If a good game has my interest and I see only the worst section which lets say last 30sec. I could assume the whole game is like this and would put me off purchasing.

if the rest of the content was in the 100s, but I knew of small areas where fps tanked. I could forgive those moments and still make the purchase. It is good to see worst case, but there should be a balanced approach.

 

[WYP]

Actually I prefer a balance. If a good game has my interest and I see only the worst section which lets say last 30sec. I could assume the whole game is like this and would put me off purchasing.

if the rest of the content was in the 100s, but I knew of small areas where fps tanked. I could forgive those moments and still make the purchase. It is good to see worst case, but there should be a balanced approach.

There is a balance to be had here and it is something that would need to be considered on a game-by-game basis.

For example, the recent Dragon Quest had tonnes of specific areas where even a GTX 1080 would dip below 60 FPS, while most game locations would allow even a GTX 1060 to achieve to sit at the game's 60FPS cap. In that case, I mentioned the dip areas in the text of the review and commented that the results and benchmarking location that I used are representative of what if a demanding area but not the game's worst case scenario.

In the case of Hitman 2, I tested the game using two benchmarking locations, one which is CPU heavy and another that is GPU heavy. The GPU heavy scene I used is very hard on Radeon GPUs, but it isn't exactly "if I look at this area just right the framerate tanks" or a "if I intentionally do this to try and make the game chugg" kinda situation. It's just how that level plays.

I have had a look at some other performance analysis pieces on Hitman 2 and it seems that Digital Foundry has spotted the same problems, as they seem to have played the game to the same mission as I have. I have also retested the RX Vega 56 at the same location and have received the same results as before.

I do plan to have a look at the new Simulation options that have been added to the game, and their impact on CPU performance. This will have to wait though, as my Church has midweek tonight.

 

[Warchild]

There is a balance to be had here and it is something that would need to be considered on a game-by-game basis.

For example, the recent Dragon Quest had tonnes of specific areas where even a GTX 1080 would dip below 60 FPS, while most game locations would allow even a GTX 1060 to achieve to sit at the game's 60FPS cap. In that case, I mentioned the dip areas in the text of the review and commented that the results and benchmarking location that I used are representative of what if a demanding area but not the game's worst case scenario.

In the case of Hitman 2, I tested the game using two benchmarking locations, one which is CPU heavy and another that is GPU heavy. The GPU heavy scene I used is very hard on Radeon GPUs, but it isn't exactly "if I look at this area just right the framerate tanks" or a "if I intentionally do this to try and make the game chugg" kinda situation. It's just how that level plays.

I have had a look at some other performance analysis pieces on Hitman 2 and it seems that Digital Foundry has spotted the same problems, as they seem to have played the game to the same mission as I have. I have also retested the RX Vega 56 at the same location and have received the same results as before.

I do plan to have a look at the new Simulation options that have been added to the game, and their impact on CPU performance. This will have to wait though, as my Church has midweek tonight.

Dont think too much into my comment. You have done nothing wrong here in Hitman. I was just stating in general, it is good to present the reviews with an overall approach. Majority of us here already have an idea of how a gpu would cope.

 

[WYP]

Dont think too much into my comment. You have done nothing wrong here in Hitman. I was just stating in general, it is good to present the reviews with an overall approach. Majority of us here already have an idea of how a gpu would cope.

My thoughts were not really directed at you, more just getting some ideas out there so that people can better understand my perspective.

 

[kardeon]

Thanks for your answer. I do understand your point of view and effectively Hitman 2 is not easy to benchmark because of the disparity of each level

 

[BFBC2033]

It all depends on the mission. Reviews have never been apples to apples even with the same hardware. Not sure why you expect anything else?

As I played the first Hitman, there were no major performance differences between all the mission, so I would assume the same for this one. And for the reviews, thats simply not true. Nearly all big reviewers get around the same performance indications in big titels (This does of course not mean the same numbers but behaviour). It is queit unusual to see such a variance.

As for the pictures they are so irrelevant that to assume it makes an article untrustworthy is to me quite childish that it bothers someone. I only mention this because you come off as an AMD guy who is upset the vega 56 and the cover article photo both show AMD falling behind.

Could you please read properly what I wrote. I never said that this review is untrustworthy, I wrote an idiom the showcase my problem with such fanboyish memes. Also this is a game review not a hardware review, its about how the game performs, not the hardware (to be more precise, a game review does not show the hardware performance, it shows the software performance, since you already know from proper benchmarks how your hardware should perform.).

I mean we already know a Vega 56 is slower than a 1070ti, so I again don't know why you expect anything else mate?

Again you deliberetly misrepresent what I said. I quoted the performance from hw's review. I even said myself that this is a bit high. I never implied or said that the vega 56 is generally faster than an 1070ti so why imply it.

Food for thought, maybe Steve's results are a bit off and not WYPs? I would be more suspect a 1070ti lost a Vega 56 than anything.
Not attacking you. Just curious as to why you think the review was bad when it honestly is one of the very few performance reviews on the internet that I found. I looked around as you suggested and found just this article. Nothing else comparing really anything(outside of consoles from Digital Foundry).

The reason why I asked WYP is because in his test every Radeon GPU performed bad, not only a few. Also the likelihood of getting an answer from him than steve is way higher.



After reading the other replies I got a good picture of how things come together. I appreciate WYP insight on his testing methodology and thoughts behind it.

 

[MacLeod]

Testing methodology aside, how is the game? I'm always in the market for a good single player action game like this. They're not making my all time favorites Batman and Dead Space. I didn't care much for Absolution tho, a little too much slow crawling and stealthing and not enough action. Is this one any better?

 

[WYP]

Testing methodology aside, how is the game? I'm always in the market for a good single player action game like this. They're not making my all time favorites Batman and Dead Space. I didn't care much for Absolution tho, a little too much slow crawling and stealthing and not enough action. Is this one any better?

I've enjoyed it. The overarching story is a bit "eh" but I have enjoyed the sandbox assassination action. There is a free trial version that you can try out, which will give you the prologue mission. It should give you an idea of what to expect.

Hitman 2 isn't really an action game, as stealth and quietly taking out targets is the goal. Agent 47 isn't great in long gunfights. The aim of the game is to search the map and discover ways to take out your target, whether it be discretely or otherwise.

It's a fun game if you want to walk around and consider every way of taking out your target, but I wouldn't describe it as action.

 

[MacLeod]

I've enjoyed it. The overarching story is a bit "eh" but I have enjoyed the sandbox assassination action. There is a free trial version that you can try out, which will give you the prologue mission. It should give you an idea of what to expect.

Hitman 2 isn't really an action game, as stealth and quietly taking out targets is the goal. Agent 47 isn't great in long gunfights. The aim of the game is to search the map and discover ways to take out your target, whether it be discretely or otherwise.

It's a fun game if you want to walk around and consider every way of taking out your target, but I wouldn't describe it as action.

Hmmm.... Well I'll probably skip it then. I don't have the attention span to handle games like that... I need some shooting and explosions lol