What to upgrade to? 670 GTX to ????

Brando457

[H]ard|Gawd
Joined
Aug 11, 2003
Messages
1,232
I bought a GTX 970 on ebay, but decided I'd rather get a 10XX card.

I have a BenQ XL2420T that does 120 hz at 1080p. I don't plan on getting a new monitor for at least a few years.

Do I get the 1060 GTX or 1070 GTX?

I'd like to max out games like GTA V, Far Cry 4, Batman etc.
 
Last edited:
At that resolution and refresh rate GTX 1070 would be my minimum choice I have the same monitor, both the XL2420T and XL2420TE (120hz and 144hz respectively) and my OC'd 980TI is not perfectly able to keep with this high refresh rate. Processor is a key part of this configuration to keep minimum frames as high as possible for an optimal experience I would look at the GTX 1080 honestly. is not easy to keep over 100FPS in most modern AAA tittles..
 
CPU: Intel I5-3570K @ 4.2 GHZ
Mobo: Asus Sabertooth Z77
Monitor: BenQ XL2420T
GFX: EVGA GTX 670 FTW
Cooling: Corsair H100I water cooling for CPU
SSD: Samsung Pro 840 128 gig for games
 
not bad at all, it's an adequate CPU but no optimal for your refresh rate, for sure you will be cpu limited even in some games unable to maintain 60fps so well.. I would still pick a GTX 1070 and pick a secondhand 3770K and overclock the hell out of it to at least 4.5ghz then you will be able to have much better performance on latest games and will be able to juice all the potential of the GTX 1070...
 
I said 3770K not 3570K.... and second hand to be cheap.. IDK what kind of budget do you have so in those case I always try to keep the budget as lower as possible.. trust me you will need all the horsepower of an i7 3770K at least at 4.5ghz to be able to sustain 60FPS in latest AAA Games.
 
i5 3570k will bottleneck the GTX1070 or 1080 little bit with those CPU intense games. there were a post about it on reddit i think , so either works fine, but as Araxis said, probably pick 1070 , and an 3770k or better
 

No, unless you can get some obscenely great price and don't expect to keep it for long. The current generation of GPUs are much better optimized for DX12. I'd rather pay a little bit more and get a GTX1070 that will perform better with future releases. Also, current-gen GPUs are much more power efficient.
 
I vote for the 1070 too. I disagree with the i5 to i7 jump though, most benchmarks show that Ivy and Sandy platforms still hold up well on most titles. I wouldn't bother trying to upgrade your CPU unless you wanted to jump to Skylake.
 
I vote for the 1070 too. I disagree with the i5 to i7 jump though, most benchmarks show that Ivy and Sandy platforms still hold up well on most titles. I wouldn't bother trying to upgrade your CPU unless you wanted to jump to Skylake.

Yeah I went GTX 1070, I am not upgrading the i5 to i7. I think people get e-boners, realistically I'm going to keep the I5 for another 3 years then upgrade mobo, cpu, monitor.
 
Yeah I went GTX 1070, I am not upgrading the i5 to i7. I think people get e-boners, realistically I'm going to keep the I5 for another 3 years then upgrade mobo, cpu, monitor.

I have a sandy bridge platform I'm pretty much planning on doing the same thing with, just 2 years or so away. My GPU is underwhelming so while I maybe bottlenecking a card like the 1070, I think it will be more up to the task when I finally upgrade the other parts.
 
not bad at all, it's an adequate CPU but no optimal for your refresh rate, for sure you will be cpu limited even in some games unable to maintain 60fps so well.. I would still pick a GTX 1070 and pick a secondhand 3770K and overclock the hell out of it to at least 4.5ghz then you will be able to have much better performance on latest games and will be able to juice all the potential of the GTX 1070...

4.5GHz is hardly guaranteed on a 3770k. That's pretty much delidding territory for that CPU unless you won the silicon+optimal IHS to core mating lottery.
 
4.5GHz is hardly guaranteed on a 3770k. That's pretty much delidding territory for that CPU unless you won the silicon+optimal IHS to core mating lottery.

what?.. nope.. if you were talking about +4.7ghz then yes.. 4.5ghz it's the least possible overclock achievable in over 20 3770K I have used ever.. with golden chips going over 4.8ghz.. if you can't get 4.5ghz in a 3770K then you don't know how to overclock... even bad overclockers 3770K can achieve 4.5ghz at 1.3v and less which it's far for needing delid, I think the worse of all 3770K ever used was able to do 4.6ghz at 1.32v without delid with just a corsair H80i as cooler so nothing great...

it's really surprising to see someone(you're the first one in fact) with a 3770K saying 4.5ghz isn't easy and it's needed delid to achieve that.. or you don't know how to overclock which it's pretty noobproof since sandy bridge or you have a faulty hardware there..
 
I vote for the 1070 too. I disagree with the i5 to i7 jump though, most benchmarks show that Ivy and Sandy platforms still hold up well on most titles. I wouldn't bother trying to upgrade your CPU unless you wanted to jump to Skylake.

I agree with this.
Ignore this CPU upgrade noise.
If you have a 1080P 120Hz display then the GTX 1070 is the way to go.
 
what?.. nope.. if you were talking about +4.7ghz then yes.. 4.5ghz it's the least possible overclock achievable in over 20 3770K I have used ever.. with golden chips going over 4.8ghz.. if you can't get 4.5ghz in a 3770K then you don't know how to overclock... even bad overclockers 3770K can achieve 4.5ghz at 1.3v and less which it's far for needing delid, I think the worse of all 3770K ever used was able to do 4.6ghz at 1.32v without delid with just a corsair H80i as cooler so nothing great...

it's really surprising to see someone(you're the first one in fact) with a 3770K saying 4.5ghz isn't easy and it's needed delid to achieve that.. or you don't know how to overclock which it's pretty noobproof since sandy bridge or you have a faulty hardware there..

Wait... You're saying if someone has a CPU that can't hit a certain OC, it's defective? Really? Can you point out where in the fine print on any of Intel's processors where they guarantee a certain overclock and thus, allow you to RMA it if it is not achievable? Because unless that's stated, you're wrong and clearly you're the type of guy who takes things personally when someone says you're wrong. Overclocking is not hard, it was a bit trick with the old i7 920, not so much since Sandy as you pointed out, so now that we get that out of the way, I must have faulty hardware, so I'd really like to see this disclaimer from Intel so I can RMA my 3770k.
 
Wait... You're saying if someone has a CPU that can't hit a certain OC, it's defective? Really? Can you point out where in the fine print on any of Intel's processors where they guarantee a certain overclock and thus, allow you to RMA it if it is not achievable? Because unless that's stated, you're wrong and clearly you're the type of guy who takes things personally when someone says you're wrong. Overclocking is not hard, it was a bit trick with the old i7 920, not so much since Sandy as you pointed out, so now that we get that out of the way, I must have faulty hardware, so I'd really like to see this disclaimer from Intel so I can RMA my 3770k.

No.. im not saying anything what you said... what I said and im going to say it again, 4.5ghz its AN EASY OVERCLOCK in any 3770k.. no matter how good or bad a 3770K can be, 4.5ghz it's far of being considered a "hardly guaranteed" achievement.. again if your 3770K can't achieve 4.5ghz you have a faulty hardware which I never mentioned to be exclusively the CPU as it could be the motherboard as in the ivy bridge time VRM was entirely dependent on motherboard (unlike haswell/devil's canyon and broadwell which was integrated on die and was removed again on skylake)...

4.5ghz it's basically a guaranteed number with crappy cooling and without delid, I certainly can't still understand your firs statement "That's pretty much delidding territory for that CPU unless you won the silicon+optimal IHS to core mating lottery." like if we were talking about 4.8ghz+ territories..

Certainly I take things personal when someone like you come saying that kind of wrong statements looking to confuse users here.. you could have said "i7 3770K at 4.5ghz it's hardly guaranteed, at least in my case I wasn't able ever to reach that speed" and good, everything perfect, I wouldn't have even replied to such commentary, but then suggesting that to achieve 4.5ghz you have to delid or win the silicon lottery because it's a totally hard achievement then yes, I can take those kind of statements as pure ignorance, I don't know what kind of experience did you have with your 3770K I don't know what kind of voltage were you expecting at overclock or what kind of stability procedures were you using, the only thing I can know certainly by your statements is that you are absolutely wrong;)...

HWbot which have a very LARGE database have the 3770K with an average of 4700MHZ ON AIR with over 47.000 submissions.. and 4600MHZ ON WATER so is not even a matter of cooling... are you going to say they are also wrong?:eek:

hwbot.PNG
 
No.. im not saying anything what you said... what I said and im going to say it again, 4.5ghz its AN EASY OVERCLOCK in any 3770k.. no matter how good or bad a 3770K can be, 4.5ghz it's far of being considered a "hardly guaranteed" achievement.. again if your 3770K can't achieve 4.5ghz you have a faulty hardware which I never mentioned to be exclusively the CPU as it could be the motherboard as in the ivy bridge time VRM was entirely dependent on motherboard (unlike haswell/devil's canyon and broadwell which was integrated on die and was removed again on skylake)...

4.5ghz it's basically a guaranteed number with crappy cooling and without delid, I certainly can't still understand your firs statement "That's pretty much delidding territory for that CPU unless you won the silicon+optimal IHS to core mating lottery." like if we were talking about 4.8ghz+ territories..

Certainly I take things personal when someone like you come saying that kind of wrong statements looking to confuse users here.. you could have said "i7 3770K at 4.5ghz it's hardly guaranteed, at least in my case I wasn't able ever to reach that speed" and good, everything perfect, I wouldn't have even replied to such commentary, but then suggesting that to achieve 4.5ghz you have to delid or win the silicon lottery because it's a totally hard achievement then yes, I can take those kind of statements as pure ignorance, I don't know what kind of experience did you have with your 3770K I don't know what kind of voltage were you expecting at overclock or what kind of stability procedures were you using, the only thing I can know certainly by your statements is that you are absolutely wrong;)...

HWbot which have a very LARGE database have the 3770K with an average of 4700MHZ ON AIR with over 47.000 submissions.. and 4600MHZ ON WATER so is not even a matter of cooling... are you going to say they are also wrong?:eek:

View attachment 5583

Sandy Bridge clocked better than Ivy and there were plenty of SB chips that would only do 4.5ghz and many more that would only do 4.2-4.4ghz....... 4.5ghz is not necessarily an easy overclock and to say someone has faulty hardware.................
 
I said 3770K not 3570K.... and second hand to be cheap.. IDK what kind of budget do you have so in those case I always try to keep the budget as lower as possible.. trust me you will need all the horsepower of an i7 3770K at least at 4.5ghz to be able to sustain 60FPS in latest AAA Games.

I'd really like to see some proof for this claim.
 
I'd really like to see some proof for this claim.

you know what you asking are in the realm of cherry picking? which is not something precisely I like but IRRC you have an ivy bridge i7 4930K right? you can test on your own disabling hypethreading and 2 cores.. anyway you asked for some games.. sadly some of the reviews are really old and show old GPUs.. however you can extrapolate and guess that with newer and much more powerful GPUs the bottleneck will be severe and also most of the time those tend to show AVERAGE FPS and not MINIMUM FPS which are the ones that have vastly difference and that's exactly the experience are going to be looking for with a 120hz+ panel the highest minimum possible FPS, but again that's something you can test and then post your own results about the experience here. you can test the games posted below and other games of your own choices..

CPUocCrysis3.png

crysis3haswell.png

CPU_2.png

CPU_01.png

CPU_01.png

CPU_03.png
 
4.5ghz it's basically a guaranteed number with crappy cooling

View attachment 5583

Is it a guarantee or not? If it is, let me see it in writing from Intel. If you can't provide that, there is no guarantee, basic or otherwise.

Sandy Bridge clocked better than Ivy and there were plenty of SB chips that would only do 4.5ghz and many more that would only do 4.2-4.4ghz....... 4.5ghz is not necessarily an easy overclock and to say someone has faulty hardware.................

Precisely this. I have a 2600k and a 3770k. My 2600k clocks better. I can get my 3770k to 4.4. Higher than that an it will throttle under IBT stress test, at least if I crank the voltage high enough for it to actually be stable and not merely "stable enough"

It's entirely possible I simply have a really bad overclocker, which just goes to show that saying 4.5 is guaranteed is a false statement.
 
It's entirely possible I simply have a really bad overclocker, which just goes to show that saying 4.5 is guaranteed is a false statement.

you have to know about CPU architecture before say sandy bridge overclocks better.. when the truly fact is due the planar architecture of sandy bridge it tend to scale better with voltages increases keeping the power at more controlable levels but also they used soldered lid which helped with the leakage produced by high temps..

Ivy bridge on the other hand due the utilization of Tri-gate 3D transistor (finfet which is a common term now due the new GPUS using it) scale less with voltages but can reach way higher clocks.. in fact most of the biggest overclocks records are with 3770K over 7ghz.. but as soon as you remove the crappy TIM you will be in the same territory as sandy bridge or even better with the use of CLU or CLP which automatically contradict with any statement about sandy bridge overclocking better.. out of the box? yes.. however the low voltages used by ivy bridge counteracts this limitation as at 4.5ghz the voltage used by a 3770K is lower than the used in a 2600K...

Again I still don't know of a ivy bridge chip unable to reach 4.5ghz... want to test?. make a new thread in overclocking & cooling section, post your entire system specs including cooler, case, airflow, ambient temp and software for stability used and then we may find why you can't reach that speed, I really invite you to that thread because I would be happy to make you contradict your own statements but also will be happy to help reach that 4.4ghz barrier and enjoy better performance in games =) because even at 4.5ghz with the 3770K I was bottlenecked in some games, it needed to me to reach 4.8ghz to remove most of the CPU bottleneck and then jump to a 4.5ghz 6700K and this is with a 980TI at 1550mhz.
 
you know what you asking are in the realm of cherry picking? which is not something precisely I like but IRRC you have an ivy bridge i7 4930K right? you can test on your own disabling hypethreading and 2 cores.. anyway you asked for some games.. sadly some of the reviews are really old and show old GPUs.. however you can extrapolate and guess that with newer and much more powerful GPUs the bottleneck will be severe and also most of the time those tend to show AVERAGE FPS and not MINIMUM FPS which are the ones that have vastly difference and that's exactly the experience are going to be looking for with a 120hz+ panel the highest minimum possible FPS, but again that's something you can test and then post your own results about the experience here. you can test the games posted below and other games of your own choices..

CPUocCrysis3.png

crysis3haswell.png

CPU_2.png

CPU_01.png

CPU_01.png

CPU_03.png

I asked for proof that

trust me you will need all the horsepower of an i7 3770K at least at 4.5ghz to be able to sustain 60FPS in latest AAA Games.
How exactly is this cherry picking? You made a pretty bold claim, so I'm simply asking you to back it up.

First two graphs = Crysis 3 but ok fine, it's a poster child for CPU scaling and actually made good use of HT after the HT patch so I expected that

Then for Dragon Age Inquisition, you show a graph of CPU benchmarks a 1280x720. I mean seriously, forcing a CPU bottleneck by running at an artificially low and unrealistic resolution isn't cool ok. If you look at the 1200p bench i5-3570K and i7-3770K perform exactly the same.

CPU_1.png

Next for RoTR, neither the i5-2500K nor the i7-3770K have issues running >60 FPS average. But ok maybe you meant MIN FPS. Fine, well comparing the i7-4770K vs i5-4690K results you see HT offers only an 8% advantage in MIN framerates, so you'd be getting 55 FPS on the i5-3570K where you'd be getting 60 on the i7-3770K. Not exactly an earth shattering difference.

Now for Fallout 4, even a dinky i3-4360 gets 60 min FPS so I'm really not sure what you're trying to say with that graph.

Lastly for Hitman Absolution, again why don't you provide the complete picture:

CPU_01.png

5% difference between i5-3470 and i7-3770K, a small amount of which could even be accounted for by the 300 MHz higher speed on the i7-3770K.

Sorry but this is hardly any proof for the statement I quoted you on. (and if I were to really nitpick, Crysis 3, Hitman Absolution, and Dragon Age Inquisition aren't exactly "latest" AAA games, well definitely not Crysis 3 or Hitman Absolution)
 
Last edited:
I asked for proof that

. How exactly is this cherry picking? You made a pretty broad claim, so I'm asking you to back it up.

First two graphs = Crysis 3 but ok fine, it's a poster child for CPU scaling and actually made good use of HT after the HT patch so I expected that

Then for Dragon Age Inquisition, you show a graph of CPU benchmarks a 1280x720. I mean seriously, forcing a CPU bottleneck by running at an artificially low and unrealistic resolution isn't cool ok. If you look at the 1200p bench i5-3570K and i7-3770K perform exactly the same.


Next for RoTR, neither the i5-2500K nor the i7-3770K have issues running >60 FPS average. But ok maybe you meant MIN FPS. Fine, well comparing the i7-4770K vs i5-4690K results you see HT offers only an 8% advantage in MIN framerates, so you'd be getting 55 FPS on the i5-3570K where you'd be getting 60 on the i7-3770K. Not exactly an earth shattering difference.

Then for Fallout 4, even a dinky i3-4360 gets 60 FPS min so I'm really not sure what you're trying to say with that graph.

Lastly for Hitman Absolution, again why don't you provide the complete picture:


5% difference between i5-3470 and i7-3770K, a small of which could even be accounted by the 300 MHz higher speed on the i7-3770K.

Sorry but this is hardly any proof for the statement I quoted you on.

Bud I base my statements on my own findings that's because I said, you have the chip? test yourself and then come here and post your findings and your experience.. because I have tested consistently across many chips including hexa and octa core chips.. you don't need proof when you have the tools to test by your own..
 
Burden of proof isn't on me since I wasn't the one making that claim.

The major difference between i5-3570K and i7-3770K is the lack of HT on the i5. You're basically saying that without HT, you simply can't get 60 min FPS, and I'm calling BS on that, because you advised the OP to spend money on a second hand 3770K when that money could be put to better use.

That's all there is.
 
Last edited:
Burden of proof isn't on me since I wasn't the one making that claim.

The major difference between i5-3570K and i7-3770K is the lack of HT on the i5. You're basically saying that without HT, you simply can't get 60 min FPS, and I'm calling BS on that, because you advised the OP to spend money on a second hand 3770K when that money could be put to better use.

That's all there is.
Well I know darn well Watch Dogs and Crysis 3 cant maintain 60 fps with HT off on my 4770k. I spent about a week playing games with HT off and several games fully pegged the cpu at times and sometimes caused some hitching even if it did stay above 60 fps. People can do whatever they want but I think a 1070 or better has no place in a system with just an i5 or less than 16 gb of system ram.
 
Burden of proof isn't on me since I wasn't the one making that claim.

The major difference between i5-3570K and i7-3770K is the lack of HT on the i5. You're basically saying that without HT, you simply can't get 60 min FPS, and I'm calling BS on that, because you advised the OP to spend money on a second hand 3770K when that money could be put to better use.

That's all there is.

You call it BS, I call you to test it for yourself... there's no better proof than anecdotal histories from the users who have the tools to take the test, is the end user who play the game for continuous hours, even major review sites only play the game for short periods of between 5 and 20 minutes, even [H] I think the longest gaming session are for 30minutes..

I said my statements based on my own sessions of gaming on my own test, you want to believe? great.. you don't want to believe? great too, I don't care but you will be the misinformed one at the end of the day, not me....
 
Well I know darn well Watch Dogs and Crysis 3 cant maintain 60 fps with HT off on my 4770k. I spent about a week playing games with HT off and several games fully pegged the cpu at times and sometimes caused some hitching even if it did stay above 60 fps. People can do whatever they want but I think a 1070 or better has no place in a system with just an i5 or less than 16 gb of system ram.

Overclocked to 4.5 GHz? Well I already said Crysis 3 was the poster child for HT scaling so no surprise there. And how many games did you test and how many games exhibited this issue for how long? If we're talking about a few hitches over a several hour gaming period that's not exactly game breaking.

You call it BS, I call you to test it for yourself... there's no better proof than anecdotal histories from the users who have the tools to take the test, is the end user who play the game for continuous hours, even major review sites only play the game for short periods of between 5 and 20 minutes, even [H] I think the longest gaming session are for 30minutes..

I said my statements based on my own sessions of gaming on my own test, you want to believe? great.. you don't want to believe? great too, I don't care but you will be the misinformed one at the end of the day, not me....

That's great. Make extraordinary claim, when asked for proof, simply say "go prove it to yourself" and walk away.
 
Back
Top