Is it time for a new gaming PC?

[MorganX]

 

I don't know. I had poor performance with RAID and when I researched it I found articles stating that the 840 Pro doesn't support TRIM in RAID mode. I was running in RAID 1 mode though. I wonder if RAID 0 is the way to go if in fact TRIM will be supported.

 

Here is what I am reading:

 

  #2  

 01-07-2011, 09:08 PM

Blue Falcon The Geek Redneck, 12.6 Years

 

Status:   

 

The Intel RST drivers have supported the ATA TRIM command since March of 2010.

It's important to note that TRIM is supported for SSDs in a NON-RAID configuration, meaning if you have multiple SSDs in a RAID configuration then the TRIM command isn't passed onto the drives.

__________________

www.bbpix.com - Free image hosting for the [H]

Intel i5-2500k @ 4.9ghz on a Gigabyte Z68XP-UD3 Mobo

16GB (4x4GB) Corsair XMS3 1600

Asus STRIX Geforce GTX 970 4GB

2x Crucial MX100 256GB (RAID0) + 1 x 2TB Seagate ST2000DM

LG 8X BD-R/RW

Corsair 750TX V2 750w PSU (Recent RMA replacement for my OG 750TX - thanks Corsair!)

"Only thing that really turns me off is odor If the vagina stinks I get turned off quick. " - rkatapt

 

 

Prior to the Series 7 chipset it did not support trim in a raid configuration. It does now.

 

If you're looking for performance, Raid 1 will not provide any redundancy, it simply mirrors the drives providing fault tolerance. Raid 0 stripes across the drives providing almost double read/write performance but provides no fault tolerance. If either drive fails you lose the data.

[patseguin Global Moderator]

OK, so I guess I'd need a new motherboard or controller card to get raid trim?

[T3X4S]

Just FYI, Intel CPUs overclock themselves nowadays, or should I say, turbo boost themselves. If you can keep it cool, you can overclock for the life of your CPU without issue. Of course it adds performance. So does buying the next highest performing CPU. Overclocking RAM is more of a hobby IMO. GPUs overclock and are sold overclocked.

 

It's the same silicon, some perform more stably and higher ratings and that's how they are binned.

So you are saying you can notice a difference between CPU A and the next one in the lineup where it is just a small clock increase ? ....  only in synthetic benchmarks - thats my point.

 

[MorganX]

So you are saying you can notice a difference between CPU A and the next one in the lineup where it is just a small clock increase ? ....  only in synthetic benchmarks - thats my point.

 

 

You would have to be specific with what CPU A and B were, but for the most part, if you have enough cooling, you can probably overclock what you have, if it's a K, as opposed to buying a new CPU.

 

The difference going from 4Ghz to 4.5Ghz, is that small or large to you.  You can notice that in general computing. Believe it or not, the noticeable performance increase from 4.5Ghz to 4.6Ghz is also noticeable. The difference between 4.6Ghz and 4.7Ghz for me, is when liquid cooling is needed as my SFF air cooled system won't run stably at 4.7 with air cooling.

 

I can do everything faster overclocked, I just don't. I can do all I want fast enough at stock speeds, but overclocking and getting the best air cooling I could has been a fun hobby/activity. Once you start its hard to stop.

 

The difference is noticeable, just not necessarily necessary and once you factor in cooking costs or in my case, noise, it may be best just to move up to the next class of processor or memory.

[T3X4S]

You would have to be specific with what CPU A and B were, but for the most part, if you have enough cooling, you can probably overclock what you have, if it's a K, as opposed to buying a new CPU.

 

The difference going from 4Ghz to 4.5Ghz, is that small or large to you.  You can notice that in general computing. Believe it or not, the noticeable performance increase from 4.5Ghz to 4.6Ghz is also noticeable. The difference between 4.6Ghz and 4.7Ghz for me, is when liquid cooling is needed as my SFF air cooled system won't run stably at 4.7 with air cooling.

 

I can do everything faster overclocked, I just don't. I can do all I want fast enough at stock speeds, but overclocking and getting the best air cooling I could has been a fun hobby/activity. Once you start its hard to stop.

 

The difference is noticeable, just not necessarily necessary and once you factor in cooking costs or in my case, noise, it may be best just to move up to the next class of processor or memory.

No its not noticeable, and no you cant notice a difference between 4.5 to 4.6 - sorry.   You might think that you notice a difference, but youre mistaken. (the difference between 4.6 and 4.7 is even less)

and no you cant do everything faster from going to 4.7 from 4.6 --

 

 

We can get into you providing evidence you think proves that someone can notice a tiny increase in clocks, but there isnt any evidence.

Going from 4.0 GHz to 4.5 might show an increase in performance of some things, but in things that I mention...things most people do all the time on their computer....nope.

[Mindovermaster Moderator]

You would have to be specific with what CPU A and B were, but for the most part, if you have enough cooling, you can probably overclock what you have, if it's a K, as opposed to buying a new CPU.

 

The difference going from 4Ghz to 4.5Ghz, is that small or large to you.  You can notice that in general computing. Believe it or not, the noticeable performance increase from 4.5Ghz to 4.6Ghz is also noticeable. The difference between 4.6Ghz and 4.7Ghz for me, is when liquid cooling is needed as my SFF air cooled system won't run stably at 4.7 with air cooling.

 

I can do everything faster overclocked, I just don't. I can do all I want fast enough at stock speeds, but overclocking and getting the best air cooling I could has been a fun hobby/activity. Once you start its hard to stop.

 

The difference is noticeable, just not necessarily necessary and once you factor in cooking costs or in my case, noise, it may be best just to move up to the next class of processor or memory.

 

I don't think he's talking about overclocking. He was comparing, at stock, 2.4 CPU1 vs 2.5 CPU2 is not that great of a gain.

[T3X4S]

I don't think he's talking about overclocking. He was comparing, at stock, 2.4 CPU1 vs 2.5 CPU2 is not that great of a gain.

There isnt a difference, true.  Simply put the CPU isnt a huge factor in overall everday performance.  An incremental clock jump of even 10% is not going to show any real world difference nowadays.

Its why benchmarks are used, its why one has to factor Pi in multiple strings - something other than real world things show the difference in OC

[cooky560 Veteran]

Build looks fine for 1080p gaming, I've seen a 770 drive a few games even at 1440p so if you must buy now, an SSD or a 770 isn't a bad bet

[Mindovermaster Moderator]

There isnt a difference, true.  Simply put the CPU isnt a huge factor in overall everday performance.  An incremental clock jump of even 10% is not going to show any real world difference nowadays.

Its why benchmarks are used, its why one has to factor Pi in multiple strings - something other than real world things show the difference in OC

 

Nothing like a switch between HDD and SSD, right?

[MorganX]

I don't think he's talking about overclocking. He was comparing, at stock, 2.4 CPU1 vs 2.5 CPU2 is not that great of a gain.

 

Oh yeah, in that case, I'd overclock. But if you're overclocking for productivity, I'd upgrade the CPU. If you already have the top end, overclock. But chances are you're doing it as an enthusiast. It's a nice hobby. Again for productivity, you've probably made more than an incremental upgrade in your workstation and don't need to overclock. For gaming, you need all you can get if you max out at hi-res over 1080p, GPU & CPU. It's a hobby not a necessity, and cheaper than SLI.

[MorganX]

There isnt a difference, true.  Simply put the CPU isnt a huge factor in overall everday performance.  An incremental clock jump of even 10% is not going to show any real world difference nowadays.

Its why benchmarks are used, its why one has to factor Pi in multiple strings - something other than real world things show the difference in OC

 

Actually, even in benchmarks overclocking CPU shows a significant difference as does CPU generations depending on the benchmark. 3D Mark uses the CPU for physics and for some reason that is a part of their scoring. CPU overclocked rigs and Haswell-E dominate based mostly on that.

[MorganX]

No its not noticeable, and no you cant notice a difference between 4.5 to 4.6 - sorry.   You might think that you notice a difference, but youre mistaken. (the difference between 4.6 and 4.7 is even less)

and no you cant do everything faster from going to 4.7 from 4.6 --

 

 

We can get into you providing evidence you think proves that someone can notice a tiny increase in clocks, but there isnt any evidence.

Going from 4.0 GHz to 4.5 might show an increase in performance of some things, but in things that I mention...things most people do all the time on their computer....nope.

 

It is and I can, and you can't prove I can't. And since you don't overclock, you're only theorizing. Not only is that .1 Mhz significant, but from 4.6 to 4.7 it's enough to make the system completely unstable or unbootable.

 

People actually use benchmarks, because it's the easiest way to "prove" it makes a difference to someone like you, not because that's the only difference it makes. Any computation that the clock speed improves in a benchmark, it will do the same for similar function during normal computing. How much of that is debatable. If you perform a particular CPU bound function in say photoshop or video transcoding, you will realize and increase. But if you do those things for a living, I would simply buy the top end Haswell-E and run at stock for stability.

[T3X4S]

It is and I can, and you can't prove I can't. And since you don't overclock, you're only theorizing. Not only is that .1 Mhz significant, but from 4.6 to 4.7 it's enough to make the system completely unstable or unbootable.

 

People actually use benchmarks, because it's the easiest way to "prove" it makes a difference to someone like you, not because that's the only difference it makes. Any computation that the clock speed improves in a benchmark, it will do the same for similar function during normal computing. How much of that is debatable. If you perform a particular CPU bound function in say photoshop or video transcoding, you will realize and increase. But if you do those things for a living, I would simply buy the top end Haswell-E and run at stock for stability.

OK - even though I dont think this is the case, I am going to be diplomatic about this and chalk it up to we aren't understanding one another.  Even though you contradict your own logic in your post - I am not gonna chase it - maybe someone else can explain it to you.  (a difference in synthetic benchmark score of 5% does not equate to a 5% performance boost ) - things like that.  Also, to prove to someone OCing doesnt really provide real world huge changes is not the reason for benchmarks.  There are 2 main reasons for benchmarks:  1.  to check overall system stability.  If it can complete a thorough stress test in a benchmark - once case assume the system is stable enough to for real world things. and 2.)  For compare/contrast with other systems/people/overclocks.  The reason they are called synthetic is because it doesnt really equate to a real world scenario... simply put - the CPU isnt a huge factor in performance.

Now if someone does a 3.0 GHz to 4.5GHz - yeah there is going to be a difference - a big one because that is a 50% increase - pretty substantial.

As for what you stated. being able to tell a performance boost from 4.6 to 4.7 - or a 2.2% increase - no you cant. - sorry.  If you felt a difference, it wasn't a CPU clock produced difference.  (I wont even get into the Law of Diminishing Returns)

As far as do I OC - yes - I used to  - Would regularly run a 20% OC - but then realized it didn't really make a difference overall and was pointless, and only potentially shortening my CPU lifespan even though I have always watercooled the processor.

 

I went and did what I said I wouldnt do - sorry for that, but I am just saying - you're mistaken, thats all, nothing personal.

 

[MorganX]

OK - even though I dont think this is the case, I am going to be diplomatic about this and chalk it up to we aren't understanding one another.  Even though you contradict your own logic in your post - I am not gonna chase it - maybe someone else can explain it to you.  (a difference in synthetic benchmark score of 5% does not equate to a 5% performance boost ) - things like that.  Also, to prove to someone OCing doesnt really provide real world huge changes is not the reason for benchmarks.  There are 2 main reasons for benchmarks:  1.  to check overall system stability.  If it can complete a thorough stress test in a benchmark - once case assume the system is stable enough to for real world things. and 2.)  For compare/contrast with other systems/people/overclocks.  The reason they are called synthetic is because it doesnt really equate to a real world scenario... simply put - the CPU isnt a huge factor in performance.

Now if someone does a 3.0 GHz to 4.5GHz - yeah there is going to be a difference - a big one because that is a 50% increase - pretty substantial.

As for what you stated. being able to tell a performance boost from 4.6 to 4.7 - or a 2.2% increase - no you cant. - sorry.  If you felt a difference, it wasn't a CPU clock produced difference.  (I wont even get into the Law of Diminishing Returns)

As far as do I OC - yes - I used to  - Would regularly run a 20% OC - but then realized it didn't really make a difference overall and was pointless, and only potentially shortening my CPU lifespan even though I have always watercooled the processor.

 

I went and did what I said I wouldnt do - sorry for that, but I am just saying - you're mistaken, thats all, nothing personal.

 

 

I'm not contradicting myself at all.

 

1) "I" never said that synthetic benchmarks equate 1:1 to real world performance. I think you said overclocking does not provide noticeable system performance increases, it only increases performance in synthetic benchmarks, and that's not true.

 

2) A lot happens to make a computer work, a CPU responds to interrupts all the time. You system buses and data transfer and processing are not saturated unless your PC and all you do is instantaneous, it is not. Your assumption is that increase in clock speed must result in 1:1 increase in all things. That simply isn't the way it works.

 

Let's say going from 4.6 to 4.7, rather 4.5 to 4.6 increases a, game, by 2 fps. That may or may not be a 2% increase, maybe significantly less. But you will notice it. And you can't think FPS, you have to think about all the additional data that has to move and physics calculations at that resolution that must take place to generate and additional 2 fps.

 

Sticking with the linear notion, if you do a thing repetitively, and the performance increases say, 1.5%, you don't think you will notice that performance increase by the end of the day.

 

I'm not a physicist so I can't explain to you all of the interaction and why that 2% clock increase may increase specific functions exponentially, but it does and or can. That is why most people use benchmarks to show the effectiveness of overclocking.

[+Anarkii Subscriber²]

OPs specs are fine, but like 90% of the posts here, I would 100% recommend a SSD even if its just to place your main OS on and keep your data drive for games. 
As for upgrading anything else, I would say hold off until about xmas this year, new chips are coming which are a vast improvement over the current gen. When they do come, get a new CPU, Mobo and Ram. Other than that, your system is pretty sweet dude. 

[MorganX]

So you are saying you can notice a difference between CPU A and the next one in the lineup where it is just a small clock increase ? ....  only in synthetic benchmarks - thats my point.

 

 

That specific statement, I can agree with. I may have missed that and only caught the general statements that overclocking only makes a difference in synthetic benchmarks.

 

Incremental step in the same CPU family, we can agree on that. Overclocking, even if you only overclock by a tenth of a Gigaherz, can make a noticeable difference. Always no, globally, no. But it can.

[+Nexus18 MVC]

PC is perfectly fine for gaming. All you need to do:

 

- overclock that 920 to 3.8-4GHZ

- get a new GPU (290/290x are stupidly cheap especially when you sell the game codes etc. on and if you prefer nvidia, the 970's are pretty good), probably wait until the new cards come out though

- maybe try to get 16GB RAM (depends on what you have running in the background etc. when gaming)

 

And depending on your future GPU, get a gsync/freesync screen.

 

You could always look at spending some money on peripherals i.e. keyboard, mouse, headset/speakers etc.

 

 

SSD's are nice but tbh there are only a few games where I really appreciate the quicker loading times i.e. BF 4

[T3X4S]

PC is perfectly fine for gaming. All you need to do:

 

- overclock that 920 to 3.8-4GHZ

- get a new GPU (290/290x are stupidly cheap especially when you sell the game codes etc. on and if you prefer nvidia, the 970's are pretty good), probably wait until the new cards come out though

- maybe try to get 16GB RAM (depends on what you have running in the background etc. when gaming)

 

And depending on your future GPU, get a gsync/freesync screen.

 

You could always look at spending some money on peripherals i.e. keyboard, mouse, headset/speakers etc.

 

 

SSD's are nice but tbh there are only a few games where I really appreciate the quicker loading times i.e. BF 4

LOL - "PC is perfectly fine for gaming"

Then you list out 5-6 things he needs to change in order to make it "fine for gaming."

The only thing you didnt list is the mobo

So which is it ?

Also, why 16GB RAM ?  He never said he is running a VM or anything that requires so much.   (not to mention its a 3x DDR3, so his RAM purchases we need to be based on 3 sticks or 6, to keep with the mobo standards (it only has 6 DIMMS)

Got to have an SSD, there is no reason not too these days.

Also, in order to do a 25% OC that you mention, he is going to need (if he doesnt have) some better cooling, and possibly a better PSU

 

Essentially, you are saying  "its fine... no problem (as long as you change 90% of the system"  

[+Nexus18 MVC]

Well his "core" setup is fine.....

 

- he doesn't need to change his hard drives and mobo

- he doesn't need to change his CPU as when overclocked, it won't hold back more powerful GPU's and he wouldn't see much, if any difference jumping to the likes of an i5 4690k etc. for purely gaming apart from a very select couple of games (which is more down to poor optimisation)

 

There are plenty of benchmarks by a number of reputable sites in various games showing a noticeable improvement in the minimum FPS when older CPU's are overclocked   Also, from my own testing I have noticed a massive difference in minimum FPS (especially in games like BF4) thus smoothness since my 290 is pretty damn powerful compared to the i5 750.

 

- I only mentioned RAM as a "possibility" because 8GB is just about enough for myself (when playing games like bf 4, my RAM usage is about 7.5GB), however, 6GB for his system might be plenty, I had 6GB before and it caused massive stuttering in BF 3, switched to 8GB, no problem at all then, number of people on overclockers.co.uk have also stated the same

 

His nvidia 760 is still "fine" for gaming but if he wants higher graphic settings with more FPS and wants to get the biggest improvement for gaming then he NEEDS to get a new GPU.

 

 

Well looking at his rig, I imagine he would have a decent enough PSU, case, cooling etc. already.... My antec 500W EA PSU coped perfectly well with an i5 750 OC to 4GHz, 7850 + 5 hard drives so he should be good to go if he has a decent quality 500W PSU. Only thing he will need to buy is a decent CPU heatsink which will only be about

[MorganX]

 

His nvidia 760 is still "fine" for gaming but if he wants higher graphic settings with more FPS and wants to get the biggest improvement for gaming then he NEEDS to get a new GPU.

 

 

 

I recently went from a 760 to a 970 and a 770 in the HTPC. The 760 I would consider mainstream but you can't max out everything at 1080p. With a 770, I'd say you could feel comfortable maxing everything at 1080p with the exception of truly demanding stuff like TressFX ... I would say 770 is serious gaming, above that and SLI hardcore, and 760 mainstream HD pc gaming.

 

I probably would not have gone 970 except the TDP vs. GPU performance was too good to resist for ~340ish. Since I don't need to go over 60fps, I don't think I'll need a GPU upgrade for long time.

[+Nexus18 MVC]

Well at the end of the day, it comes down to what FPS people are happy with, for myself, I need 50+ at all times, anything less and it just feels horrible imo.

 

A 970 compared to the 760 is quite a bit of an upgrade:

 

http://www.anandtech.com/bench/product/1038?vs=1355