|
It would cost more to make 2 separate dies, but they do have CPUs with the GPUs cut, the 2550k. It's not like the GPU's wasting a significant amount of power when it's disabled either.
|
#
¿
Mar 9, 2012 00:10
|
|
|
|
| # ¿ Apr 26, 2024 03:17 |
|
|
Removing an IHS is trivial and free so anyone who cares about overclocking can do it. They're just cutting costs by making assembly simpler.
|
|
#
¿
Apr 28, 2012 07:20
|
|
|
Alereon posted:I would hardly call this trivial, and this is on a CPU that's already dead so the guy didn't need to worry about damaging it: https://www.youtube.com/watch?v=lMzzUuvKWPM That's going way overkill, you don't need to cut the epoxy all around the edges. Just to make sure I wasn't misremembering it I just went out and popped off an IHS from a P4. I tried to record a video showing how easy it was, but it looked like garbage because I needed 2 hands to stabilize the CPU. The entire effort was under 45 seconds though. All you need to do is push the IHS a little bit to the side so it breaks the bond and then it falls right off. The hardest part is figuring out how to clamp it so you can exert pressure evenly on the side of the PCB, but just on a corner of the IHS. Edit: I watched it and it's actually not as incomprehensible as I imagined. I cheated by putting the CPU in a loose socket I had so I wouldn't need to cushion the back edge with a piece of cloth, but that's just so it's easy to see.https://www.youtube.com/watch?v=AfEDateV6ik craig588 fucked around with this message at 09:41 on Apr 28, 2012 |
|
#
¿
Apr 28, 2012 09:16
|
|
|
3570s using the stock cooler.
|
|
#
¿
Jun 2, 2013 00:42
|
|
|
Shaocaholica posted:I think it's quite genius/douchey that intel is using tim under the ihs. As long as the CPUs are still within an acceptable thermal envelope, there's no need to do what would amount to helping overclockers. It also basically makes voiding your warranty the cost of admission for -real- over clocking. Just like in the good old days only now it's glaringly detectable. Most importantly it means the CPUs cost less to manufacture.
|
|
#
¿
Jun 3, 2013 18:19
|
|
|
Yeah, I've always done the vice+block method to remove IHSes and by comparison the razor way looks tedious and nerve racking. I made a video a while back showing that it takes literally seconds once you have the CPU clamped.
|
|
#
¿
Jun 9, 2013 02:14
|
|
|
I actually didn't know it ever was safe to put a computer to sleep with a USB drive plugged in, I always just assumed something unexpected might happen.
|
|
#
¿
Jun 10, 2013 20:27
|
|
|
Probably. Realtemp reads retail C2Ds fine, I actually was using it on a E8400 just a couple days ago. Engineering sample CPUs can potentially be on incredibly early or even one off silicon that's not 100% finished.
|
|
#
¿
Jun 29, 2013 03:37
|
|
|
P.N.T.M. posted:I read somewhere (maybe this thread), that the "core ix" series is an evolution of the Pentium III architecture. That was the original Core series sometime around 2005-2006, before they added the "ix" part to the Core brand. They were kind of like an extremely highly clocked Pentium 3 with a lot of cache, expanded instruction sets, and multiple cores.
|
|
#
¿
Oct 23, 2013 23:19
|
|
|
I used to use a laptop with an Intel integrated GPU and a similar CPU, though a 2.3GHz 133MHz (533MT) FSB version, and it handled 720P h264 at 3500kbs easily using Media Player Classic. I don't think 1080P would be entirely unreasonable, depending on the bitrate. Youtube at even 720P might be a stretch though.
craig588 fucked around with this message at 19:10 on Feb 15, 2014 |
|
#
¿
Feb 15, 2014 19:08
|
|
|
They don't fall much because everyone has figured out when to stop producing old stuff so supply dries up at retail right when they're ready to ship their next generation. It's usually best to buy right when a generation shifts. Usually price drops only happen when there's a surprise from competition and they're forced to respond. Unless you're looking at used, there will be tons of people upgrading and selling their old stuff for 50-100 off retail.
|
|
#
¿
Jul 17, 2015 21:04
|
|
|
The extra 2 cores and 64GB of memory are more of a benefit than the extra lanes. If you're not going to need 6 cores or 64GB of memory and can't afford 980tis with a 5820k I'd shuffle everything around to a 4790k and 980tis. Even with only 4 lanes videocards still aren't running into PCIE bandwidth limits.
|
|
#
¿
Sep 16, 2015 17:22
|
|
|
It's held true since they were first introduced: basically anything without heatspreaders will be the best ram you can get, both in terms of cost, but also in terms of long term reliability. Performance has been meaningless since the latter part of the DDR2 era, so you can get whatever's cheapest from any name brand you like.
|
|
#
¿
Oct 25, 2015 04:09
|
|
|
Berious posted:I'm really agonizing over upgrading my CPU. Have a MSI MS-7756 h97 Mobo with an i5-3570 (no K) CPU.. DDR4 memory is within like 20 dollars of DDR3 memory, occasionally it even ends up cheaper for DDR4 memory, depending on sales and sizes you need. DDR4 dropped out of early adopter price territory really quickly.
|
|
#
¿
Dec 4, 2015 18:56
|
|
|
I'm happy with my Notcua D15 with my 5820K. It's way quieter than the Corsair H100 and H110 I've tried and is good for 4.3Ghz below 80C or 4.7Ghz if I was comfortable getting up to 90C. That testing was done with the newest prime95 running over 24 hours with AVX2 and FMA instructions which I found was able to generate the most heat. Processor variation YMMV and all.
|
|
#
¿
May 4, 2016 09:33
|
|
|
You'll need a C series chipset motherboard (expensive) but not because of the ES nature. the EP processors are made to be used in dual socket configurations and single socket motherboard compatibility isn't guaranteed, though some manufacturers have figured out workarounds to get them running on X79 chipsets. This is probably a better bargain if you're looking for a cheap highly threaded CPU transcoding box, around 100 dollars for 8 cores and around 24GB of memory in a complete system just add drives. http://www.natex.us/product-p/fquanta98l.htm
|
|
#
¿
Jun 2, 2016 17:01
|
|
|
If you can wait, wait. Skylake has higher IPC and overclocks better than Haswell and Broadwell. There will always be something new coming out in a year though, if you can use 6 cores now, get a 5820k. I had to get a new PC last month and felt like I had reliable leaks on the overclockability and pricing so I bought a 5820K and I'm glad I didn't wait now that official Broadwell-E reviews are out. If you're mostly playing games, get a 6600K.
|
|
#
¿
Jun 2, 2016 23:31
|
|
|
lDDQD posted:that Xeon that turbos to 5.1GHz out of the box Was a fake or something so specialized for a high end customer Intel didn't even bother putting it on ARK.
|
|
#
¿
Jun 3, 2016 16:40
|
|
|
If you're crazy about noise like me it's still nearly as quiet as a D15 while at slightly slower clock speeds. I recently gave away an old 2500K machine and in doing do I had to replace the D15 with a 212 so I could move the D15 to my new machine. I went from 4.6 on the D15 to 4.3 on the 212 for around the same temperatures and noise level. A very good heatsink all around as long as you're not trying for the limits of silence or clock speed, the only thing to watch out for is it's famously annoying to install, but you do that once and then it sits there forever.
|
|
#
¿
Jun 5, 2016 22:37
|
|
|
The H110 GT vs GTX are just two different manufacturers, the GTX has some sort of colored lighting while I think while people seem happier with the long term reliability of the GT, they perform about identically if I remember right. As someone who has tried the H100, H100i and a H110 and went back to a D15, if you have the space in your case for it, look into a Noctua D15 with some upgraded (up to) 3000 RPM PWM fans. It's much quieter than the H110 in normal usage and can cool much better for stress testing when I ramp up the fans. Currently using a overclocked 5820k with a D15 with 3000 RPM temperature controlled fans that I never hear move above 800 RPM in anything but prime95, including video encoding and games simultaneously.
|
|
#
¿
Jun 6, 2016 23:42
|
|
|
Motherboard makers have the option of locking down GPU overclocking even if it's supposed to be supported by the chipset and processor. If you have an OEM motherboard I'd bet they don't allow GPU overclocking. I ran into a similar problem when I wanted to try overclocking an otherwise great laptop that was held back by its thermally limited GPU so if anyone happens to know a workaround I'd love to know it, but from my research mobo makers can deny you access to the GPU clock controls.
craig588 fucked around with this message at 00:31 on Jun 26, 2016 |
|
#
¿
Jun 26, 2016 00:28
|
|
|
wipeout posted:Apart from core clocks, what are the advantages of raising cache / uncore clocks on a 6600k? Worth it? Virtually nothing, some extra heat and potential stability issues, an extra GHz might get you 2% more performance.
|
|
#
¿
Jul 4, 2016 22:52
|
|
|
Yeah, I quickly searched to see if you posted about your videocard, and didn't find a post, but if you have anything less than a 1080 you should be looking at a videocard upgrade before your CPU. And if you already have a reasonably modern videocard it might be worth looking into software issues first. Even 20% seems extremely optimistic for a jump from a 4590 to a 6700K, I'd expect it to be more in the 5-10% range.
|
|
#
¿
Jul 31, 2016 10:45
|
|
|
You might be right there, it depends on the motherboard. With more overclocking focused ones you can set it to ignore load power limits, but with a non overclocking 4690 he probably has a non overclocking motherboard too. I'm not loyal to any particular site so school me on benchmarks if these are bad, but these are always the sort of graphs I think of when I think about CPU upgrades for gaming. http://www.anandtech.com/show/9483/intel-skylake-review-6700k-6600k-ddr4-ddr3-ipc-6th-generation/16 a couple games can use the extra CPU performance, and most of them hardly notice even going way back to a Sandybridge. craig588 fucked around with this message at 10:58 on Jul 31, 2016 |
|
#
¿
Jul 31, 2016 10:54
|
|
|
Tthe 6600K is on sale too and the better option for most people mainly interested in playing games http://www.monoprice.com/product?p_id=16383 212.49
|
|
#
¿
Sep 24, 2016 02:57
|
|
|
All the way up through a QX9650 I was one of those advocates for never using power saving features for extra stability. Starting with Sandybridge (I ended up skipping the 1366 generation) I left everything on because it didn't feel like it was compromising anything anymore. Had a 4.6Ghz Sandybridge and now a 4.7Ghz Haswell. They keep getting better and better about switching states stabily and I think a lot of people who don't use power saving remember the pre Sandybridge era where you were potentially giving up a lot of performance. On the far far end with sub ambient cooling it could be a different story, but for anyone with air or water I think it's mostly lingering thoughts from old power saving working pretty badly than anything based on real world testing of recent hardware.
|
|
#
¿
Oct 31, 2016 15:18
|
|
|
BIG HEADLINE posted:EDIT: Oh, and the 5820K can only support 64GB of memory while the 6800K can support 128GB. Figured I'd mention that. Not true. The 5820K can handle 128GB. Unless there's some bug where it'll report 128GB, but only actually use 64GB of it? I'm only using 16GB of memory because the type of stuff I do doesn't demand memory at all so I don't have first hand experience, but all of the documentation says a max of 128GB. I probably could have gotten away with 8GB, but I'm not building a PC with 8GB of memory in 2016.
|
|
#
¿
Nov 4, 2016 14:05
|
|
|
That's similar to what I did with my 5820K. I could get 4.7Ghz (probably) stable, but it was touching 90C under prime so I backed off to 4.3 because that keeps it under 70C and I haven't noticed the performance hit. I wasn't doing it so much for the power use, but for the hopeful longevity of keeping it 6+ years by being cooler, lower power use just goes together with that. At one time I'd think that was crazy, who'd ever want to use a 5 year old processor? but I kept my Sandybridge for 5 years only replacing it because I had to replace another old computer that physically failed and figured I might as well buy new.
|
|
#
¿
Nov 23, 2016 01:41
|
|
|
Once you switch to a K, add in the Z motherboard and even just a 212 cooler you're looking at about a 100 dollar premium over a non overclocking setup. If you're not planning on overclocking it's a waste.
|
|
#
¿
Nov 26, 2016 04:34
|
|
|
Probably not, but if it's real competition it will at least force Intel to lower prices a bit.
|
|
#
¿
Nov 27, 2016 21:40
|
|
|
Good luck getting a 3770K that'd hit even 4.6GHz. At 4.6GHz you'd be lucky to tie the performance of a 4.8GHz Sandybridge, forget about beating it. I really don't think you should be very CPU limited with a 4.8GHz Sandybridge, I went from a 4.6GHz Sandybridge to a 4.7GHz Haswell, but I keep it at only 4.3GHz for the temperature benefits (doesn't break 70C during Prime) and it feels like it performs the same as the 4.6 Sandybridge as to being no issue at all to game performance. I got the Haswell replacement because I needed another computer and might as well buy something new and conga the old PC down. I think any upgrade would put you into a very similar performance place and high CPU times probably are a coding problem with the game rather than something that could be easily solved with more processor power. The best reason for enthusiasts with highly clocked Sandybridges to get new processors now is when we want to get an additional PC, there's no sense in buying old stuff when you need more PCs, but it's still real hard to justify processor upgrades.
|
|
#
¿
Dec 10, 2016 05:23
|
|
|
Harik posted:CPU-Z is even more confused: Shuffle your memory sticks around, you're running in single channel. I wouldn't worry about the timings at all (you could always manually set them if you want, but eh) but dual channel is very important.
|
|
#
¿
Dec 12, 2016 20:10
|
|
|
Guru3d is all over the place because they have some real knowledgeable writers and also have some people who seem to not know english very well and no editor to fix anything up.
|
|
#
¿
Dec 14, 2016 04:51
|
|
|
I thought it performed slightly worse than the 1060 and wasn't available for MSRP for months. I remember a long period where you could get 200 dollar 1060s and 480s were up at 240 dollars. People speculated that the harder to find 4GB model was just for a bullet point 200 dollar price point and 480s actually would stay at 240 (which proved not to be true, but that was how hard it was to find a 200 dollar 480).
|
|
#
¿
Jan 2, 2017 08:46
|
|
|
EdEddnEddy posted:I wonder how many more 2500K/2600K's will be ran until they actually die vs any other chip? I mean really when has the CPU actually been the cause of a new system rebuild vs almost any other part around it. I replaced a 4.6Ghz 2500K with a Haswell not because of the Sandybridge failing, but because of my server machine based on a 10 year old C2Q failing. Figured it'd be one of the only good chances I'd get to upgrade my main computer so I turned the Sandybridge into the new server machine.
|
|
#
¿
Jan 17, 2017 12:26
|
|
|
The biggest problem you'll run into with Windows 9x is lack of processor clock speed support. There are many layers of official and unofficial patches you can go through to get stability up to around 2.4Ghz. Out of the box you're looking at a 350Mhz limit and with official patches up to 2.1Ghz. That's just for the OS, there's still software that wasn't expecting clock speeds that fast and will crash on you anyways. There's a whole world of people building new machines to run legacy software on. For drives people have all sorts of solutions, what I've most frequently gone with is compact flash. It's compatible with IDE and systems with IDE drives weren't expecting more than even a few MB/s and often are limited to a 33MB/s bus that never *really* was supposed to be fully saturated so even a 50MB/s CF card ends up exposing never before seen bottlenecks and stability issues from reading data too fast.
|
|
#
¿
Feb 4, 2017 19:18
|
|
|
I can't find a live version of the KB article now, but on Google you'll find a lot of dead links to the 350MHz patch Microsoft released. Here's the article about the 2.1GHz limit https://support.microsoft.com/en-us/help/312108/windows-protection-error-in-ndis-with-a-cpu-that-is-faster-than-2.1-ghz
|
|
#
¿
Feb 5, 2017 15:12
|
|
|
Don't buy anything name brand from aliexpress. They're great if you want something fake and cheap, but you're just as likely to get a board with replaced caps from a no name brand if you're buying a "new" one from aliexpress. Replaced sockets are also a pretty common thing to do, people break the pins in them and if you have a reflow machine you could revive a board for a few dollars, but same deal that you're probably getting something not actually new with unknown history.
|
|
#
¿
Feb 16, 2017 23:48
|
|
|
It's supposed to be to allow people to use older wiring that isn't up to 10g specs. I don't think there has been enough real world use yet to see how effective that proves to be or if everyone needs to upgrade their old cat 5 anyways if they have runs longer than like 10 feet or something.
|
|
#
¿
Feb 18, 2017 18:15
|
|
|
|
| # ¿ Apr 26, 2024 03:17 |
|
|
I got caught in the fake cache PC Chips and SIS board scam http://redhill.net.au/b/b-bad.html. Since then I've always been too scared of cheap motherboards to ever go cheap on them and I've never had a problem from any of Asuses high end to mid range boards. I'm at a point where I'm so familiar with Asus terminology and layout that I'll pay more for an Asus board just so I don't need to think about how settings are renamed or rearranged between manufacturers. An extra 30 dollars up front is worth it so I don't have a chance of making a mistake and wasting hours because I missed a setting. Asrock spun out from Asus. They started out as a budget line but engineers couldn't leave things alone and eventually they were making competitive enthusiast boards.
|
|
#
¿
Feb 23, 2017 18:36
|
|