Batman: Arkham Knight will require flagship GPU for Ultra quality

[Guest]

I'm not saying that what you've said is untrue - but don't use passmark to try to make a point. That site is pretty crap. I'm sick right now and too tired to look into the details of what you've said so I don't even know if I agree with you... but please don't use passmark.

Care to provide a source to explain and back up the claim that Passmark's CPU benchmarks are "pretty crap"? What exactly makes it bad?

 

[Lionvibez]

No, it does not.
For starters, "quad-core CPU" is a completely meaningless term. Are you suggesting a $50 Athlon 5350 (a quad-core Jaguar CPU) will run this game well? Or are you suggesting a Core i3 will not?
That's specially significant because that chart lists the Phenom II X4 965 and the Core i5-750 as a minimum requirement, but a dual-core Core i3-4130 is faster than both, according to passmark.
My money is on Core i3 CPUs giving at the very least 80% of the performance of a Core i5 CPU, and Core i7 ones providing no benefits over a Core i5, like it always is. And I wouldn't be surprised if an overclocked Pentium G3258 also runs this game reasonably well, despite having two cores and no hyper-threading.

I believe you are wrong.

A dual core may run this game but how well.
A hyperthreading core does not = a full core.

An i3 or pentium G3258 will not be as fast as a full quad core. We will have to wait for Techspot performance review but my gut tell me a quad will be required.

 

[Lionvibez]

Care to provide a source to explain and back up the claim that Passmark's CPU benchmarks are "pretty crap"? What exactly makes it bad?

Its a synthetic benchmark that's the reason.

 

[Guest]

An i3 or pentium G3258 will not be as fast as a full quad core. We will have to wait for Techspot performance review but my gut tell me a quad will be required.

Did you even see the article, right here on TechSpot, comparing the performance of Core i3, Core i5 and Core i7 CPUs in games? If you didn't, here: https://www.techspot.com/review/972-intel-core-i3-vs-i5-vs-i7/
The conclusion is that with the GTX 980, a Core i5 already offers only a small benefit over a Core i3 for almost twice the price. With a GTX 960, that difference is practically null.
It's very safe to say this game, as well as any game this year, will run perfectly well on a Core i3 CPU.

Its a synthetic benchmark that's the reason.

Every CPU benchmark is a sythetic benchmark. That does not mean it produces irrelevant results. It tests the same tasks on different CPUs and offers an average of all submission results. It's a perfectly fine way to compare CPUs, and could be done with any other comprehensive benchmark, like PCMark (only not with specific benchmarks that measure one task, like Sandra or SuperPI).
Unless you provide a reason why Passmark specifically is an inefficient benchmark, this argument is invalid. "All synthetic benchmarks suck" is not the answer.

 

[Lionvibez]

I've seen the review and while its good and up to techspot high quality.
I don't form an opinion off one review. When this game is out and a full performance review is out with those dual cores in the lineup then we shall see if your claims are true.

And I'm not someone that puts alot of weight in synthetic benchmarks so while for you its may be valid for me it is not. We can agree to disagree on the worth of synthetics.

 

[Guest]

I've seen the review and while its good and up to techspot high quality.
I don't form an opinion off one review. When this game is out and a full performance review is out with those dual cores in the lineup then we shall see if your claims are true.
And I'm not someone that puts alot of weight in synthetic benchmarks so while for you its may be valid for me it is not. We can agree to disagree on the worth of synthetics.

Oh, but you don't need to form an opinion off of one review (even though several games were reviewed in that article). You can search for the TechSpot's performance reviews of every major game release in the last few years. You can search for Core i3 CPU reviews on several other technology sites. You'll see that Core i3 CPUs always, consistently perform very close to Core i5 CPUs in discrete GPU gaming, in EVERY review you can find.
As for synthetic benchmarks, you don't have to care for any of them. The fact here is simply that the Core i3-4130 is a faster CPU than both the Core i5-750 and the Phenom II X4 965. So it's completely absurd to think that those processors will run the game and the Core i3 won't, based solely on number of cores.

 

[Peter Farkas]

No, it does not.
For starters, "quad-core CPU" is a completely meaningless term. Are you suggesting a $50 Athlon 5350 (a quad-core Jaguar CPU) will run this game well? Or are you suggesting a Core i3 will not?
That's specially significant because that chart lists the Phenom II X4 965 and the Core i5-750 as a minimum requirement, but a dual-core Core i3-4130 is faster than both, according to passmark.
My money is on Core i3 CPUs giving at the very least 80% of the performance of a Core i5 CPU, and Core i7 ones providing no benefits over a Core i5, like it always is. And I wouldn't be surprised if an overclocked Pentium G3258 also runs this game reasonably well, despite having two cores and no hyper-threading.

Yeah you are right, I agree with the above. I didn't mean any Athlon by quad core obiously. Those days are gone...
But simply the fact that they mention a quead core CPU as minimum req, means that the game is developed to utilize 4 threads, and in these scenarios an i5 performs significantly better than an i3 which can only handle 4 threads through HT. i3 falls behind pretty much in such games. That's why I said I am keen to see the perf benchmark results.
Passmark is just a synthetic benchmark which is not a solid basis for comparing gaming performance.

 

[Guest]

Yeah you are right, I agree with the above. I didn't mean any Athlon by quad core obiously. Those days are gone...
But simply the fact that they mention a quead core CPU as minimum req, means that the game is developed to utilize 4 threads, and in these scenarios an i5 performs significantly better than an i3 which can only handle 4 threads through HT. i3 falls behind pretty much in such games. That's why I said I am keen to see the perf benchmark results.
Passmark is just a synthetic benchmark which is not a solid basis for comparing gaming performance.

That's exactly the same thing people were saying about Watch Dogs, Far Cry 4, and GTA V, all of which lists quad-core CPUs as "minimum requirements". But then, predictably:
- https://www.techspot.com/review/827-watch-dogs-benchmarks/page5.html
- https://www.techspot.com/review/917-far-cry-4-benchmarks/page5.html
- https://www.techspot.com/review/991-gta-5-pc-benchmarks/page6.html
Core i3 CPUs run all those games perfectly well. In the worst scenario for the i3, which was GTA V, it still provides 84% of the performance of an i5- 4690K. And it was above 60 FPS too.

i3 falls behind pretty much in such games.

So no, it absolutely does not.
As for Passmark, I'm not using it to compare gaming performance. I'm using it to state that in general computing, the i3-4130 is faster than the i5-750 and PII X4 965. Therefore, thinking the latter can run a game while the former, which is faster, can't, is completely absurd.

 

[Lionvibez]

I would like to see the numbers of the i5-750 overclocked at 4Ghz vs the i3 overclocked.

I bet you the gen 1 i5 will produce a better gaming experience.

 

[Peter Farkas]

That's exactly the same thing people were saying about Watch Dogs, Far Cry 4, and GTA V, all of which lists quad-core CPUs as "minimum requirements". But then, predictably:
- https://www.techspot.com/review/827-watch-dogs-benchmarks/page5.html
- https://www.techspot.com/review/917-far-cry-4-benchmarks/page5.html
- https://www.techspot.com/review/991-gta-5-pc-benchmarks/page6.html
Core i3 CPUs run all those games perfectly well. In the worst scenario for the i3, which was GTA V, it still provides 84% of the performance of an i5- 4690K. And it was above 60 FPS too.

So no, it absolutely does not.
As for Passmark, I'm not using it to compare gaming performance. I'm using it to state that in general computing, the i3-4130 is faster than the i5-750 and PII X4 965. Therefore, thinking the latter can run a game while the former, which is faster, can't, is completely absurd.

Yeah, those specs detail quad core cpu as minimum reqs but can't truly make advantage of it.

While there are some games where even an i5 easily beats the i3 badly:
https://www.techspot.com/review/320-civilization-v-performance/page12.html
This is also true for the Metro series, Dragon Age, etc...

The i3 is a very good entry level gaming CPU but don't forget that your GPU also determines what CPU you should have, otherwise you can run into CPU bottlenecks. In case of the i3 you can have CPU bottleneck if you have anything stronger than a GTX960 based on the below article:
https://www.techspot.com/review/972-intel-core-i3-vs-i5-vs-i7/

So yeah, the i3 is a sweet spot for gamers, but you have clear benefits from an i5 or i7 over the i3 if your pocket doesn't bother. I didn't say the i3 can't run those games so I don't know where you got that...
I personally wait for the benchmarks to see how it really performs.

 

[Kevin82485]

Most retailers have a no questions tasked return policy within a week - you could get a 290 or a 290x for around $50-70 more than you paid for your card.

You usually have at least 10 days to return a product for free. Go get yourself a proper 290X with 4 GBs of VRAM or 8. Either that, or pony up for a 980 or SLI/Crossfire setup.

I have a G-Sync monitor so I gotta stick with nVidia. Really though the 970 is perfectly ok with me. The jump in performance relative to the cost of a 980 just is not worth it to me.

Also, I bought the card from Newegg, and as much as I love their prices, their returns process kinda sucks. Any time I have had a return I gotta pay for return shipping, and I have to pay bogus restocking fees. I only buy from them now if I am 100% certain of my decision. I would buy from Amazon instead, but in my state Amazon charges me sales tax whereas Newegg does not, so I save some additional money through Newegg that way. But Amazon's return process is much better in my opinion. Damned if you do, damned if you don't I guess.

 

[Lionvibez]

So yeah, the i3 is a sweet spot for gamers, but you have clear benefits from an i5 or i7 over the i3 if your pocket doesn't bother.

I would go so far as to say anyone buying a Dual core cpu now for a gaming rig in 2015 is seriously misguided or just an *****. At least try to future proof your rig alittle if possible.

 

[Peter Farkas]

I would go so far as to say anyone buying a Dual core cpu now for a gaming rig in 2015 is seriously misguided or just an *****. At least try to future proof your rig alittle if possible.

totally agreed. Espcially if you have a decent GPU it is a must to have at least an i5

 

[Guest]

I would like to see the numbers of the i5-750 overclocked at 4Ghz vs the i3 overclocked.
I bet you the gen 1 i5 will produce a better gaming experience.

What a coincidence, you can!
https://www.techspot.com/review/379-crysis-2-performance/page8.html
This is before the Sandy Bridge days, so there isn't a more recent i3 in the mix. But you can see the i5-750 being insignificantly faster than a Nehalem/Westmere-based i3.
https://www.techspot.com/review/467-skyrim-performance/page7.html
Here you can see the Sandy Bridge i3 outperforming the i5-750.
https://www.techspot.com/review/458-battlefield-3-performance/page7.html
Again, Sandy Bridge i3 and i5-750 performing exactly the same.
Since then, the i5-750 has been removed from the benchmarks. But since we've seen about 20~25% improvement in IPC between Sandy Bridge and Haswell, as well as a sight clock bump between the 2120 and the 4130, it's not hard to imagine what it would look like today. Also, you can go back to any performance test in the last few years and still see i3 processors consistently beating Phenom II X4 models (still in the benchmarks), which is the other "quad-core" recommendation made by this game.
Also, the i5-750 is not overclockable (just like the i3). The "K" models already existed back then, you'd need a much more expensive i7-870K to overclock.

Yeah, those specs detail quad core cpu as minimum reqs but can't truly make advantage of it.
While there are some games where even an i5 easily beats the i3 badly:
https://www.techspot.com/review/320-civilization-v-performance/page12.html
This is also true for the Metro series, Dragon Age, etc...
The i3 is a very good entry level gaming CPU but don't forget that your GPU also determines what CPU you should have, otherwise you can run into CPU bottlenecks. In case of the i3 you can have CPU bottleneck if you have anything stronger than a GTX960 based on the below article:
https://www.techspot.com/review/972-intel-core-i3-vs-i5-vs-i7/
So yeah, the i3 is a sweet spot for gamers, but you have clear benefits from an i5 or i7 over the i3 if your pocket doesn't bother. I didn't say the i3 can't run those games so I don't know where you got that...
I personally wait for the benchmarks to see how it really performs.

The Civilization review, while a fair example of i5s being significantly faster than i3s back then, it from the Nehalem/Westmere days. It's an entirely different thing to compare a Nehalem i5 to a Haswell i3.
And you must be having some trouble comprehending the "i3 vs i5 vs i7" article. In it, when using the GTX 980, the only games where the difference between an i3 and an i5 is significant is in Metro Redux (about 18%), and arguably in Thief (about 11%). It's <10% in every other case, and again: with the GTX 980. For a CPU that costs just 53% of the price of the i5 (i3-4130 vs. i5-4690), so even with the GTX 980 the i3 is still much better value. Keep in mind that on both Metro Redux and Thief the i3 still stays above 60 FPS.
Now, with the GTX 960, that difference disappears completely. The worst case is again Metro Redux, and the i3 provides a bit over 93% of the performance for about 53% of the price. This is the worst case for the i3.
Finally, the article makes it very clear that there are no tangible benefits in getting an i7 for gaming. Best case scenario is 8% higher performance than the i5 in Metro Redux. I don't know where you got that idea from.

I would go so far as to say anyone buying a Dual core cpu now for a gaming rig in 2015 is seriously misguided or just an *****. At least try to future proof your rig alittle if possible.

totally agreed. Espcially if you have a decent GPU it is a must to have at least an i5

Objectively wrong.
Someone buying an i3 in 2015 gets a much better value than any quad-core CPU and mostly insignificant reductions in performance, as show in the comparison article. That same article also shows that an i5 provides almost no benefit for a "decent" GPU like the GTX 960/R9 285, and for high-end GPUs (you didn't mean to imply only high-end GPUs are "decent", did you?) like the GTX 980 or R9 290X an i5 can bring benefits, but an i3 can run it perfectly well too, so an i5 is certainly not a "must have".
Of course, if you have the money, then by all means, get the i5. But the whole point is that saying things like "this game requires a quad-core", "the i3 falls behind pretty much in such games", "having an i5 is a must" and "poeple who don't buy quad-cores are *****s" are objectively and demonstrably wrong statements.

 

[Peter Farkas]

What a coincidence, you can!
https://www.techspot.com/review/379-crysis-2-performance/page8.html
This is before the Sandy Bridge days, so there isn't a more recent i3 in the mix. But you can see the i5-750 being insignificantly faster than a Nehalem/Westmere-based i3.
https://www.techspot.com/review/467-skyrim-performance/page7.html
Here you can see the Sandy Bridge i3 outperforming the i5-750.
https://www.techspot.com/review/458-battlefield-3-performance/page7.html
Again, Sandy Bridge i3 and i5-750 performing exactly the same.
Since then, the i5-750 has been removed from the benchmarks. But since we've seen about 20~25% improvement in IPC between Sandy Bridge and Haswell, as well as a sight clock bump between the 2120 and the 4130, it's not hard to imagine what it would look like today. Also, you can go back to any performance test in the last few years and still see i3 processors consistently beating Phenom II X4 models (still in the benchmarks), which is the other "quad-core" recommendation made by this game.
Also, the i5-750 is not overclockable (just like the i3). The "K" models already existed back then, you'd need a much more expensive i7-870K to overclock.


The Civilization review, while a fair example of i5s being significantly faster than i3s back then, it from the Nehalem/Westmere days. It's an entirely different thing to compare a Nehalem i5 to a Haswell i3.
And you must be having some trouble comprehending the "i3 vs i5 vs i7" article. In it, when using the GTX 980, the only games where the difference between an i3 and an i5 is significant is in Metro Redux (about 18%), and arguably in Thief (about 11%). It's <10% in every other case, and again: with the GTX 980. For a CPU that costs just 53% of the price of the i5 (i3-4130 vs. i5-4690), so even with the GTX 980 the i3 is still much better value. Keep in mind that on both Metro Redux and Thief the i3 still stays above 60 FPS.
Now, with the GTX 960, that difference disappears completely. The worst case is again Metro Redux, and the i3 provides a bit over 93% of the performance for about 53% of the price. This is the worst case for the i3.
Finally, the article makes it very clear that there are no tangible benefits in getting an i7 for gaming. Best case scenario is 8% higher performance than the i5 in Metro Redux. I don't know where you got that idea from.



Objectively wrong.
Someone buying an i3 in 2015 gets a much better value than any quad-core CPU and mostly insignificant reductions in performance, as show in the comparison article. That same article also shows that an i5 provides almost no benefit for a "decent" GPU like the GTX 960/R9 285, and for high-end GPUs (you didn't mean to imply only high-end GPUs are "decent", did you?) like the GTX 980 or R9 290X an i5 can bring benefits, but an i3 can run it perfectly well too, so an i5 is certainly not a "must have".
Of course, if you have the money, then by all means, get the i5. But the whole point is that saying things like "this game requires a quad-core", "the i3 falls behind pretty much in such games", "having an i5 is a must" and "poeple who don't buy quad-cores are *****s" are objectively and demonstrably wrong statements.

Who are you little princess?
though I have to admit I agree with most of what you said. ;-)

 

[Lionvibez]

What a coincidence, you can!
https://www.techspot.com/review/379-crysis-2-performance/page8.html
This is before the Sandy Bridge days, so there isn't a more recent i3 in the mix. But you can see the i5-750 being insignificantly faster than a Nehalem/Westmere-based i3.
https://www.techspot.com/review/467-skyrim-performance/page7.html
Here you can see the Sandy Bridge i3 outperforming the i5-750.
https://www.techspot.com/review/458-battlefield-3-performance/page7.html
Again, Sandy Bridge i3 and i5-750 performing exactly the same.
Since then, the i5-750 has been removed from the benchmarks. But since we've seen about 20~25% improvement in IPC between Sandy Bridge and Haswell, as well as a sight clock bump between the 2120 and the 4130, it's not hard to imagine what it would look like today. Also, you can go back to any performance test in the last few years and still see i3 processors consistently beating Phenom II X4 models (still in the benchmarks), which is the other "quad-core" recommendation made by this game.
Also, the i5-750 is not overclockable (just like the i3). The "K" models already existed back then, you'd need a much more expensive i7-870K to overclock.

I said overclocked numbers all those links are for stock clocks.

Battlefield 3 on a 64 player map will choke on dual core chip. I've seen in person and there are no amount of graphs you can link to that will show other wise because nobody post numbers from BF multi on their reviews. Most of the battlefield numbers you see on the net are from the single player portion of the game.

Here are crysis 2 cpu scaling numbers from a different site.
http://www.tomshardware.com/reviews/crysis-2-directx-11-performance,2983-8.html

Skyrim runs well on most processors so not the best example to prove your point.

Until I see benchmark numbers for this game on a i3 which we have to wait so you still haven't proven anything

And I will say it again anyone buying a dual core to build a gaming rig with in 2015 is an *****.
The extra money you spend on the quad will be justified in the system lasting your longer, not to mention even gamers do more on their pc than just play games. Any kinda of multitasking while said game is running will be alot better on a quad.

And you saying the i5-750 is not overclockable is wrong.
Pre K series intel chips you overclock with blk I still know people running this chip at 3.8 or 4Ghz.

K series just made it easier for noobs to overclock because you are just changing the multiplier.

I've been overclocking cpu's since the early 1990's.

 

[Guest]

I said overclocked numbers all those links are for stock clocks.
(...)
And you saying the i5-750 is not overclockable is wrong.
Pre K series intel chips you overclock with blk I still know people running this chip at 3.8 or 4Ghz.
K series just made it easier for noobs to overclock because you are just changing the multiplier.

Fair enough, I stand corrected about that. But the point remains, the recommended specs here is a "Core i5-750", not an "overclocked Core i5-750". Since current Core i3 processors perform better than a stock i5-750, it's safe to infer that they will run this game too.
The same applies to Battlefield. If a Core i3-4130 suffers on BF multiplayer, then a stock i5-750 will just as well, because the four 2.66 GHz Nehalem cores on the i5 are slower in general than the two 3.4 GHz Haswell cores with HT on the i3-4130. It's even worse if the game only uses two threads, because then it would become two 2.66 GHz Nehalem cores vs. two 3.4 GHz Haswell cores, leaving the i3 with an obvious advantage. Bottom line: if the i3-4130 suffers, so does the i5-750. And if the i5-750 runs it well, then so does the i3-4130.

Skyrim runs well on most processors so not the best example to prove your point.

Skyrim runs well on most processors today. Not all that much back in 2011, as you can see in the performance review. And that's exactly the point, if games made extensive use of multiple cores, Core i5 and i7 processors would be significantly ahead of the i3. The problem is exactly that they don't, which is why an i3 performs so close to the i5 and i7 ones (and why an i7 has practically no advantage over an i5). We're all waiting for this mythical day when multi-core CPUs will be heavily used in games, but it's been seven years since Core 2 Quad CPUs became mainstream and it still never happened.

Any kinda of multitasking while said game is running will be alot better on a quad.

Multitasking is a fair argument in favor of the i5. But it's not a necessity, and doesn't mean that this game (or any for that matter) requires an i5 or won't run on an i3.

Finally, that Crysis 2 scaling isn't very much relevant to the i3 vs. i5 discussion, because it looks at core scaling on a Phenom, with no HT/SMT. It's one thing to run a game in two cores and two threads, and another to run a game with two cores and four threads. A Haswell processor with HT wouldn't scale the same way that Phenom did.

 

[CoolGuyJ]

Is there any chance at all that I can run this with my rig?

CPU: Intel Core i7-3610QM, 2.3 GHz
GPU: Nvidia Geforce GT 630M v2 (2GB)
RAM: 8 GB
HDD: 1 TB
Native resolution: 1920 x 1080

I know I have a huge bottleneck with my GPU, but Shadow of Mordor had (insane) GTX 460 as its minimum requirement, and I can comfortably run that game in 1366x 768 with 22 ~ 40 fps (depending on area). [Honestly, Shadow of Mordor had such extreme requirements, but such stupendous optimization...]

I know 30fps + is the minimum fps most people expect, but I honestly have no trouble with fps as long as it's 20 or above, and have no qualms in lowering my resolution (as long as I can get playable fps of 20+)

So considering the above in mind, is there chance I can run it (if optimized well), or is there no hope at all?

 

[Kevin82485]

Is there any chance at all that I can run this with my rig?

CPU: Intel Core i7-3610QM, 2.3 GHz
GPU: Nvidia Geforce GT 630M v2 (2GB)
RAM: 8 GB
HDD: 1 TB
Native resolution: 1920 x 1080

I know I have a huge bottleneck with my GPU, but Shadow of Mordor had (insane) GTX 460 as its minimum requirement, and I can comfortably run that game in 1366x 768 with 22 ~ 40 fps (depending on area). [Honestly, Shadow of Mordor had such extreme requirements, but such stupendous optimization...]

I know 30fps + is the minimum fps most people expect, but I honestly have no trouble with fps as long as it's 20 or above, and have no qualms in lowering my resolution (as long as I can get playable fps of 20+)

So considering the above in mind, is there chance I can run it (if optimized well), or is there no hope at all?

I wouldn't count on it playing very smooth. It's tough to guess though. How does Batman Arkham Origins (or any of the previous titles in the series) play on your laptop? Arkham Knight is going to be even more resource hungry than before. If Arkham Origins (or others) runs between 20 - 30 FPS, which you said you are ok with, I don't think there is anyway possible that Arkham Knight will get to even 20 FPS.

 

[game474]

I'm at the recommended specs on my rig so won't have any issues there.

I just need to finish batman AK city before I pick this up.

I totally skipped origins as I heard it wasn't that great and had a few game stopping bugs.

No way on hell man. I couldn't finish AC because of the DrFreeze crash. But I played Origins. I know the ign ratings and whatnot are bad. But that game surpasses AC by miles in terms of gameplay and graphics. And the story: Black Mask announces a million dollars for an assassin who can kill batman. Where does Joker come into this? Awesome twists. And everything in the game is much more difficult (I think I played on Normal only). Even ordinary fights with thugs is tough. The boss fights are insanely designed and quite time consuming and difficult. The game really looked beautiful on my bravia!

 

[Guest]

Will an i5 4690k suffice for ultra

 

[Kevin82485]

Will an i5 4690k suffice for ultra

Depends on your video card(s)

 

[Peter Farkas]

Depends on your video card(s)

Should be enough from a CPU prespective

 

[scorpian007]

Passmark is the worst way to determine gaming performance, passmark is optimised to utilise the cores and threads to their maximum capability. Gaming is different, I'd say you'd need a Core i5 level quad-core CPU to run this game smoothly.

 

[Peter Farkas]

Passmark is the worst way to determine gaming performance, passmark is optimised to utilise the cores and threads to their maximum capability. Gaming is different, I'd say you'd need a Core i5 level quad-core CPU to run this game smoothly.

What CPUs do you mean by Core i5 level quad-core CPUs beside the i5 itself?