Core i7-4790K ‘Devil’s Canyon’ overclocking revisited

Our first attempt at overclocking one of Intel’s new “Devil’s Canyon” processors was, frankly, a little bit underwhelming. Intel had pitched these new processors as especially good for high clock speeds, thanks to some changes to the heat-transfer and power-delivery bits in the CPU package, but we couldn’t get our Core i7-4790K review sample to run any faster than a regular Haswell-based 4770K. Many others in the press saw similar results. We liked other things about Devil’s Canyon, such as the 4790K’s higher stock clock speeds at the same price as the 4770K, but its overclocking prowess just didn’t impress.

Then, frustratingly, we heard whispers from one vocal Intel employee who suggested that final, production versions of Devil’s Canyon might perform better. I’ve gotta say, as a reviewer, I’m not fond of that entire concept. My great hope and expectation is that whatever chip we’re reviewing is pretty well representative of what people will be buying soon.

But whatever. Maybe, in this case, it was true. Maybe there would be something different about the chips that hit store shelves. There is, of course, one simple way to find out.

That’s why I was particularly receptive when the folks at MSI extended a tantalizing offer: let us send you a retail version of the Core i7-4790K along with one of our fancy Z97 MPower motherboards. You can overclock it and see how it goes.

Well, okay, then. Seems like it’s worth a shot.

We’re just looking at a single sample of a retail Core i7-4790K, so we’re not exactly doing science to this thing. Whether or not our results will track with your own, should you buy a 4790K, is pretty much anyone’s guess.

Heck, I didn’t expect this, but MSI shipped the Core i7-4790K to us in the box with the Z97 MPower motherboard, covered in nothing but bubble wrap. I was somehow expecting a retail boxed processor, with the ain’t-been-opened stickers still intact. Perhaps I should have mentioned that to them. The markings on the CPU’s metal cap indicate a final product, but for all I know, MSI went through a big stack of 4790K chips and sent only the most frequency-friendly examples to guys like me. So caveat your emptors, folks. Who knows what any of this means.

I’ve got the MPower

I was happy to see the Z97 MPower make its way into Damage Labs. At $185 online, this is the sort of high-end board that, frankly, we’ve kind of avoided at times here at TR, simply because we’re ridiculously frugal and tend to focus our reviews on cheaper boards where possible. The Z97 MPower is part of MSI’s overclocking-centric motherboard series, and you’ll pay a bit more to get its suite of extra-fancy features.

I’ve gotta admit, some of the perks are nice, not least of which is the look of the board itself. The MPower may be the handsomest motherboard ever to grace Damage Labs with its presence. Many of the premium features of the MPower’s spec sheet are incredibly nebulous things like “enhanced components” and “OC certified” that don’t make a lot of sense to me. (Who does this OC certification? Is there a board with strict criteria? Somehow, I doubt it.) Still, MSI has included some obvious nods to hard-core overclockers, like the row of multimeter-ready voltage-check terminals next to the ATX power connector and the ability to clock DDR3 DIMMs at speeds up to 3300 MT/s.

And, hey, this is just a really nice Z97 mobo, with a slot for M.2 SSDs and a gaggle of SATA 6Gbps and USB 3.0 ports.

The Z97 MPower’s firmware interface is very similar to the one in MSI’s Z97 Gaming 7 board that Geoff reviewed not long ago. You’ll find everything you need there to make an Intel processor do unnatural things with its clock frequencies, and the interface itself looks as pretty as, you know, motherboard firmware can probably get.

I have to admit that I struggled to find a few settings every now and then, in part because I’ve spent more time recently with boards from Gigabyte and Asus. MSI has made some nice progress on its firmware, but the organization is a little different. For instance, the pane on the right side of the main menu where one can access the fan-speed controller disappears entirely once you start modifying something else, like CPU voltage. To get it back, you have to press the Escape key one more time than my instincts tell me should be necessary. Once you find the appropriate menu, though, the fan-control interface is slick and capable.

I did run into a few honest-to-goodness rough edges with the MPower board, too. For one thing, the CPU fan headers on this mobo only support PWM-style four-pin fans, a fact that isn’t spelled out terribly prominently in MSI’s documentation. The DC fan on my Thermaltake NIC C5 cooler acted strangely, and I couldn’t sort out the problem until I happened to mention it to Geoff. He quickly explained that he’d run into the PWM limitation on some of MSI’s other boards, and I instantly knew what was up. If you don’t retain a Geoff on staff like I do, sorting out this problem may take you longer.

Fortunately, the MPower’s system fan headers do support DC-style fans, and their speed-control settings can be set to key off of CPU temperatures. In seconds, I had the NIC C5’s fan ramping up and down smoothly in response to CPU temperatures.

 

In search of higher frequencies

These days, amazingly enough, motherboards will overclock your processor for you. I’ve been aware of this fact for a while, especially since Geoff has been testing auto-overclocking features in motherboard reviews, but I’ve never actually bothered to try them. Not for years, at least, since those features were in their infancy and didn’t do an especially good job.

You see, I have mad skillz involving multiplier and core voltage settings. I can change both variables at once in the BIOS, yo. F10 and it’s on, friend.

But I figured this time around, I’d give MSI’s OC Genie button a try. Seems like a gimmick, putting a push-button right on the motherboard itself that says “overclock this thing,” but the actual concept appears to be sound in theory. MSI maintains a set of profiles for common CPUs and will invoke a fitting profile when you press the OC Genie button. What you’ll get probably won’t be the highest possible speed for your processor, but it should be a reasonably decent overclock with a high probability of success. Why not start there?

I hit the OC Genie button, powered on the system, and BAM! Error message. Overclocking failed. Press F1 for the BIOS menu and whatever.

I made several more attempts, but no dice. The board threw an error every time I tried to boot with OC Genie enabled. I’m not sure what exactly the problem was, because this feature is kind of a black box. Were the memory timings too aggressive? Did it need a little more CPU voltage? Impossible to say.

Being the curmudgeon that I am, I brushed off this problem with a shrug and went to work in the firmware menus myself. Who needs a genie, anyhow? I’ve got this.

To be clear, what I’ve got is a really simple approach to overclocking an unlocked CPU: strap on a big-ass tower cooler, crank up the multiplier to something plausible, and raise the CPU core voltage as needed. Repeat the mult and voltage tweaks until something breaks.

With a little coaxing, I was able to get the retail Core i7-4790K running and reasonably stable at 4.8GHz and 1.32V.

I tried for 4.9GHz, and the chip was able to boot into Windows at that speed, but it quickly crashed when I started up our Prime95 load test. I tried raising the voltage in steps, going as high as 1.424V, but the chip just wasn’t happy at that clock speed.

4.8GHz at 1.32V ain’t bad. We were able to benchmark our pre-release sample of the 4790K at 4.7GHz, but it needed 1.45V to get there, which is quite a bit of juice for a 22-nm processor. And that earlier 4790K was a little iffy at 4.7GHz. Given everything, I’d call this retail chip 200MHz better, if you’re looking for a realistic overclock for long-term use. (Those retail CPU voltages, by the way, come from MSI’s monitoring software. I found that if you asked the board for 1.3V in the firmware, you’d get 1.32V. Ask for 1.35V, you get 1.368V. And so on. It errs on the side of more juice, which is a little scary when you’re hitting the limits of your chip or cooler.)

That clock speed improvement isn’t gonna set the world on fire, and this one sample isn’t enough to establish any sort of a trend. Still, every data point we’ve seen points to something in the range of 4.7 to 4.8GHz being the most common limit for a Devil’s Canyon CPU—which is awfully similar to a regular Haswell K-series part.

And, you know, it’s really fast.

With the retail CPU at 4.8GHz and running Prime95, our Z97 MPower-based test rig drew about 210W under load at the wall socket. CPU temperatures peaked at about 80°C. The system’s power use dropped down to about 54W at idle on the Windows desktop, which is almost exactly what Geoff saw from the Z97 Gaming 7 with a stock-clocked, default-voltage processor.

It may well be possible to squeeze even higher clock speeds out of a Devil’s Canyon CPU with extreme cooling and the like, but that is a question for the dudes with liquid helium pots. I am not one of those dudes.

That said, I’m gonna geek out over these things in my own way soon, with broader performance testing of a whole range of the latest CPUs. That’s a task we’ve neglected for too long around here, and I intend to rectify the oversight. Stay tuned for more CPU coverage coming up shortly.

Comments closed
    • LoneWolf15
    • 5 years ago

    Like Chrispy, I’m really wondering (well, I think he’s wondering) how an i7-2600k in the overclocked range compares.

    Mine runs at stock voltage at 4.2GHz all day long, under load. It might go higher; I just decided not to push it. I have no doubt it beats a stock 3770k; how does it stand up against a 4790k, at least at stock?

    I keep getting tempted to buy one of these 4790k chips the next time I’m at a MicroCenter (where they are $60 cheaper than NewEgg/Amazon), but I just don’t know if I can justify it when the time comes.

    • oldDummy
    • 5 years ago

    [quote<]...That said, I'm gonna geek out over these things in my own way soon, with broader performance testing of a whole range of the latest CPUs.....[/quote<] Thing is there hasn't been a bunch of movement regarding top end CPU's. my i7-3970x runs two cores at 4.6Ghz rest at 4.4 in turbo mode. Runs toasty with 80C+ common at full loads [3.9Ghz] utilizing 100H AIO. my point: While this cpu is around two years old there isn't many that can whip it for all around functionality. Your testing will give all a benchmark, that can't help but be a good thing. Looking forward to it.

    • elmopuddy
    • 5 years ago

    I just changed the multiplier in the bios from 44 to 47, and set “sync all cores”.. perfectly stable.

    This is with the CM 280mm AIO cooler, Asus Sabertooth Z97

    • Kougar
    • 5 years ago

    Thanks for the update, interesting tidbit of info!

    Regards to that CPU comparison testing ya mentioned, I’d love to see a Q6600 and i7 920 thrown in!

    • flip-mode
    • 5 years ago

    Honestly, overclocking doesn’t look much worth the effort at all. Performance at stock speed is right on the heels of performance at your max overclock.

      • derFunkenstein
      • 5 years ago

      But that’s because stock speed is 500MHz faster than the 4770. If this thing came in at 3.6Ghz stock and 4.7Ghz OCd the difference would be wider.

        • Firestarter
        • 5 years ago

        So, buy an i7-4790K and leave it at stock. It seems that this is the best option for a lot of enthusiasts out there, but it does sound counterintuitive for a processor that is supposed to be the ultimate in overclocking performance!

          • derFunkenstein
          • 5 years ago

          Maybe a little, but I’d take another 500MHz on top if I can leave it at stock and get it to run at 4.5.

        • flip-mode
        • 5 years ago

        Right – I should ninja edit to say that I’m not complaining – I’m do not criticize the fact that this processor is stock-clocked up to the point that overclocking it is much less appealing. I think it’s great, to be honest. If work needs any computers built any time soon this will be the processor that goes into them.

    • spugm1r3
    • 5 years ago

    Out of curiousity, what BIOS did the board have on it? I wouldn’t suspect that MSI would send a board to illustrate their OC Genie without ensuring that it functioned well first (cough… that NEVER happens in industry…), but I do know that they released a newer BIOS at the end of June to cope with overclocking the refresh Haswells.

      • Damage
      • 5 years ago

      Latest one from late June, downloaded from their website.

        • EzioAs
        • 5 years ago

        I’m not sure on the details but maybe the OC Genie didn’t work is because the 4790K is already overclocked reasonably well? I mean, it’s basically a 4770K with higher stock speeds.

          • Damage
          • 5 years ago

          Could be a memory timings setting. Could be anything. No way to tell!

          Honestly, it was not something I was gonna lose sleep over. Doing it yourself is way better.

            • EzioAs
            • 5 years ago

            Of course, I was just just trying to find the reason why the auto OC didn’t work. Not that I know anyone who uses such a thing anyway.

            • nanoflower
            • 5 years ago

            I used it with my G3258. It was a quick and easy way to get the system up and running while OCed without spending a lot of time finding that edge of stability. So I think it’s useful for people who either don’t know how to OC or just don’t have the time. In my case it defaulted to a 4.3GHz speed with 1.3 volts which proved stable through a 1 hour run of prime 95. I’ve since been able to improve on those results.

            I’m not sure why it didn’t work for Scott but as he said there could be a number of reasons it didn’t pass. Without knowing what MSI defaults to when using OCgenie it’s hard to say.

            • Milo Burke
            • 5 years ago

            I hate to be the voice of dissent, but I am interested in auto-overclocking. Overclocking is a bit confusing to me, and it comes with some risk of frying the thing. I want some free performance, but if I’m going to wrap my mind around some new complex thing, it might be options trading instead of overclocking.

            Perhaps TR could post an overclocking guide that is as helpful and easy to understand as the System Guide consistently is.

            Or, instead, perhaps TR could post something like:

            [quote<]If you have the 4790k, a decent overclock most of you can get with an acceptable tower cooler is XX multiplier and X.XX voltage. And here are some decent numbers for your 4690k. And let's not forget your Pentium Anniversary Edition. If you have a more obscure processor, sorry we can't help. Check back after Broadwell desktop chips come out for the Broadwell equivalents. And if you're feeling more adventurous, dive deeper into overclocking with this helpful guide at some other site.[/quote<]

            • derFunkenstein
            • 5 years ago

            For me, it was just a matter of cranking the multiplier until the stock voltage couldn’t keep it stable, then dialing it back a notch or two. With socket 1155 or 1150, it’s just a matter of setting the turbo speeds to what you want them to be and rebooting to see if it works. It’ll still slink back down to the slow speeds when it’s idle and crank it up when it’s busy.

            Unfortunately, suggesting XX multiplier and X.XX voltage for any model of chip is a risk, because some may not make it.

    • tbone8ty
    • 5 years ago

    please add a fx-9590 to your testing

      • Deanjo
      • 5 years ago

      Why when we know that it gets slaughtered already from lower end, generation old intel chips?

      • Chrispy_
      • 5 years ago

      What, so we can see that it’s fighting i3’s whilst drawing double the power?

        • sschaem
        • 5 years ago

        Probably way more then double the power.
        But if you look at any ‘modern’ workload, the i3 got no chance to match the fx-9590.

        side note: the 4.7ghz vishera is rated at 220w, While the 4.ghz model is rated at 125w
        So in short almost 100w to go up by 700mhz…. Thats insanely stupid.

        But it work both ways. And there is a sweet spot for power efficiency on those chips.
        (4.7ghz is not it)

        Personal experience:

        I see the FX-8350 for $149 from time to time, and my experiment show you can drop the TDP to <100w without down clocking, and go much lower if you are willing to set the base at 3.8ghz.
        (I would do this if I was stuck using the insanely loud stock cooler)
        sucks that their is no decent motherboard available, making the CPU non interesting for a new PC.

        But I did endup recently with a fx-8320 based workstation. And settled with at 4.2ghz 1.225v
        (could go higher, but I wanted to keep the max TDP low so it would be a silent PC)

        The reason for this is that AMD provide its CPU profiling tools for free, its about $700 for vtune if I went with an i7 based system.

        Its not a speed monster, but it does deliver about 115% the performance of a stock i7-3770k when I run some of my code (ray tracing type stuff. mimic very closely povray results actually)

        For reference, my second monitor (a 24″ dell CFL) actually use more power then this PC when doing code editing. So that was eye opening. So super happy in term of the system power usage and noise under load.

    • DPete27
    • 5 years ago

    [quote<]The MPower may be the handsomest motherboard ever to grace Damage Labs with its presence.[/quote<] My mommy says "It's what's on the inside that matters."

    • Chrispy_
    • 5 years ago

    [quote<]And, you know, it's really fast.[/quote<] Unless you've had a 2600K for the last three and a half years. Those generally overclock better than Ivy or Haswell, or Devil's canyon and Intel's IPC has barely changed over the last two generations. The only way you could call it really fast is if you limit everything to stock speeds, in which case 4GHz/4.4GHz on boost is pretty sweet.

      • HisDivineOrder
      • 5 years ago

      Intel IPC has improved, though.

      SB to IVB, 7% maybe?
      IVB to Haswell, 10% probably.

      Assuming you could boost your overclocking to 4.7-4.8ghz and you have a built-in IPC advantage along with chipset improvements that came with Z87 or Z97 compared to SB especially (especially related to PCIe and SATA), I’d say you might see more of an improvement than you think.

      That’s assuming you meet or beat your SB overclock.

      Now if you’re arguing that the SB was a great deal given how long it’s remained relevant? Uncontested. Especially if you bought a 7970 way back when they were first introduced (or when they dropped in price due to the arrival of the GF 680) because that’s a system that’s remained relevant a very, very long time.

      It’s just you shouldn’t delude yourself into thinking you wouldn’t see a decent-sized performance advantage in a number of important areas because you would. Is it worth the investment cost to go from SB to Devil’s Canyon? Eh… To some, I suppose.

      If it were me, I’d wait for Haswell-E.

        • Chrispy_
        • 5 years ago

        This is my beef; The amount of cherry-picking and effort going into overclocking these 4790K’s is ridiculous.

        This was obviously not just any retail-boxed 4790K because the glaring omission of the retail box might as well have been a giant-neon sign. It’s obviously 200MHz faster than tthe typical samples.

        Cherry-picked 2600K’s were hitting over 5GHz with AIO coolers three and a half years ago, and 7% IPC from Ivy, plus 10% IPC from haswell means that the 4790K has an 18% advantage using your numbers. If we spent this much effort picking through a pile of 2600K chips and tweaking the voltage, 2600K we’d have 5.2GHz maybe more. Just google 2600K 5GHz club there are thousands of people who have acheived 5GHz, and plenty of those are over 5.2GHz on plain water loops.

        So, 4.8GHz x 1.18 (for the IPC advantage) is 5.6GHz near enough. Sandy was doing close to this [b<]42 months ago.[/b<] Is this what we call progress these days? As I said, the only real advantage of Devil's Canyon is that the stock clocks are up for those people who don't have Z-series boards. 18% more IPC and the 600MHz base-clock advantage is actually something you [i<]can[/i<] notice.

        • CrazyElf
        • 5 years ago

        [quote<] SB to IVB, 7% maybe? IVB to Haswell, 10% probably. [/quote<] It's much smaller than that. More like: Ivy, 3-5% Haswell, it varies. Things that could not use the new instruction sets 5%. Things that could, 20%. The only reason to upgrade is if you have something that can take advantage of the latest and greatest instruction sets. Quite a few Sandy Bridge chips were able to do 5GHz at under 1.45V and remain stable (Intel Burn Test stable). The 22nm chips seem to be capping at 4.7-4.8 GHz, which has remained true for both Ivy and Haswell Devil's Canyon even factoring in the delid.

      • derFunkenstein
      • 5 years ago

      People with Sandy Bridge got in on the ground floor of something amazing and should stay there for sure.

    • Meadows
    • 5 years ago

    At first I was surprised by the more casual language, but then I saw the results.

    The improvement in voltage requirement looks very good though.

    • Deanjo
    • 5 years ago

    [quote<]We were able to benchmark our pre-release sample of the 4790K at 4.7GHz, but it needed 1.45V to get there, which is quite a bit of juice for a 22-nm processor. And that earlier 4790K was a little iffy at 4.7GHz. [/quote<] Just for clarification, you were using the same MSI board with both the pre-release and retail version of the 4790K or are you comparing what you could do with the board that you used to test the pre-release 4790k?

    • drsauced
    • 5 years ago

    I just finished putting together an i7-4790K for my work machine this afternoon. The old machine was having, uh, *cough* hard drive problems *cough*. Chief observations, it must run very cool, I spotted the heatsink fan actually stopping for short periods of time with the Z97-A mobo. The CPU hovers around 34C most of the time with a Thermalright U120 HSF. Yeah, yeah, it’s a BIOS tweak issue.

    Anyway, I haven’t taxed it at all, but WS2008R2 installed on the MX100 512GB SSD in about 18 minutes. Surprisingly, it isn’t as fast booting the same OS as my home i5-3570K with a trusty, and rather good, Samsung 830 256GB SSD. But the new machine is very fast otherwise.

    The main problem was with Thunderbird and the search function chewing up the HDD. Both the 16GiB of RAM and SSD ought to bring the 40+ minute wait in the morning to tolerable levels. The old WD 2TB drive was having some bad sector issues (so no, the upgrade is not entirely unjustified 🙂 )

    • Sargent Duck
    • 5 years ago

    So…how do I go about getting my own Geoff?

      • Wirko
      • 5 years ago

      There’s one spare Jeff over at TR; you might want to ask him if he’s into extreme overclocking on air.

      • Neutronbeam
      • 5 years ago

      Rent-a-Geoff?

    • NeelyCam
    • 5 years ago

    I wonder if this improved retail goodness extends to i5-4690K…

      • nanoflower
      • 5 years ago

      It should given that other people are seeing similar results with the 4790K and the kind of results people are getting with the cut down G3258.

    • Bensam123
    • 5 years ago

    Not to keep poking, but what were the Load Line Calibration settings and do you guys adjust them? Just leaving things on auto doesn’t really tell us what level they were on or even if they were on. It’s a pretty important setting to overclocking stability and I’m not saying that to be a elitist jerk or anything. It really helps with stability and most people who do a decent amount of OCing can confirm this. Just set it for 75% or equivalent setting and give it a go.

      • f0d
      • 5 years ago

      im pretty sure they know what they are doing and they adjusted the LLC

      also i bet they adjusted the VTTCPU and CPUPLL (or equivalents) voltages also for a more stable overclock

        • Bensam123
        • 5 years ago

        Those matter much less then LLC. AFAIK they just tweak the multipliers and voltage, but not LLC… which rates right up there with vcore if you’re trying to get a decent OC.

          • f0d
          • 5 years ago

          oh i agree they matter less but that doesnt mean that TR doesnt adjust them also

          i always thought they adjusted everything

    • Krogoth
    • 5 years ago

    The 22nm process is just leaky. It also needs the volts to keep itself stable at 5Ghz and beyond. It was engineered for low power consumption, while increasing the transistor budget. It wasn’t meant for super-high clockspeeds.

    It is hard to say if Broadwell 14nm will suffer the same fate.

    • MadManOriginal
    • 5 years ago

    I’m not sure I’d even bother overclocking one of these if I got it. The stock clocks are pretty awesome, too bad they didn’t do the same with the 4c/4t chip….maybe I’d just set it to run at the default max turbo for any load and just be happy.

    • nanoflower
    • 5 years ago

    Yes, with liquid nitrogen you can do some amazing things. I saw one report over on hwbot of a guy getting over 6GHz on a Pentium G3258 with liquid nitrogen. That’s truly amazing.

      • Deanjo
      • 5 years ago

      [quote<]I saw one report over on hwbot of a guy getting over 6GHz on a Pentium G3258 with liquid nitrogen. That's truly amazing.[/quote<] and truly useless for any practical purpose.

        • f0d
        • 5 years ago

        drag cars are useless for practical purposes also but that doesnt mean that they are not any fun

          • Deanjo
          • 5 years ago

          Drag cars actually provide entertainment and are a product of some talented engineering teams, mechanics and also test human physical capabilities. They would be boring if they just calculated the theoretical times and threw them up on a board.

        • nanoflower
        • 5 years ago

        No one suggested that people are going to run a typical system with liquid nitrogen. Scott mentioned that you could likely overclock even more with extreme cooling so I mentioned an example of such cooling. I certainly never suggested it’s anything other than an example of what extreme work can accomplish. It’s certainly nothing I’m likely to ever attempt even for a test, but I don’t have anything against the people out there discovering just what is possible.

          • Deanjo
          • 5 years ago

          I don’t even consider it discovering “what is possible” if it cannot be sustained with real world use. It’s about as useful a test as slapping some JATO rockets on a bike to see if it can fly.

    • chuckula
    • 5 years ago

    [quote<]With the retail CPU at 4.8GHz and running Prime95, our Z97 MPower-based test rig drew about 210W under load at the wall socket. CPU temperatures peaked at about 80°C.[/quote<] Personal experience with my delidded 4770K + a Kraken X-60 cooler (280x140 mm radiator) is that I can get the hottest cores up into the low 90C range using Intel's linpack torture test on the right workloads (full AVX2 in use). I'm clocking up to a maximum of 4.7GHz, so you've done pretty well hitting 4.8GHz with lower temps and no need for delidding. One crazy idea though.... there are apparently some MSI motherboards that have a special bracket to brace the cooling block against the bare die instead of the heatspreader... so if you were to delid you could theoretically just put your cooling block right on the die. Now I'm not quite crazy enough to try that... BUT I CAN ALWAYS ASK/BEG YOU GUYS TO DO IT!!!

      • USAFTW
      • 5 years ago

      You delidded your 4770K? Shot of the delidded chip and your balls of steel next to it or it didn’t happen.

        • chuckula
        • 5 years ago

        I don’t know if the flickr account still works but I posted detailed information on the whole process here: [url<]https://techreport.com/forums/viewtopic.php?f=2&t=88253&hilit=lived+to+tell[/url<]

          • USAFTW
          • 5 years ago

          Wow, I saw that thread, didn’t know it was yours. Great job. Not something I would try with my own though, not that lucky. A friend tried to delid the CELL inside a 2007-era PS3 and destroyed it in the process.

Pin It on Pinterest

Share This