I got 3x E6500K in today. Someone said I should be concerned a bout fakes. The vendor didn't bother to clean them. 3 different levels of dirt on them. Veeeeerrrrrrrry clever. Fake Chinese dirt!
I'm very happy with my TJ258 XPS720 heatsink purchase, For $20 I was expecting used, but it's NOS in the box!
I see I'm not the only one then that keeps an eye on the better Dell heatsinks. Whenever I see a good one for a low price, I'm tempted to buy it. Good heatsinks are always worthwhile getting if they are going cheap, even if they are proprietary and don't fit standard stuff. I regularly adapt my own mounting mechanisms for custom heatsinks on GPUs.
I did a little more testing 4.3Ghz 4 core isn't going to happen.
4.15 is good though. That's the "normal" overclock for these.
That still seems insane to me (in a good way ).
It's essentially a 50+% performance boost.
I wonder how a 4+ GHz Core 2 Quad architecture compare to something like an i5-2500 or a similarly clocked 1st get i5/i7 with same core and thread count (so I guess no i5's then, lol ).
I still haven't upgraded the 2x E5606 CPUs in my Precision T7500, and looks like TS or SetFSB can't do anything to OC. Nonetheless, 2 x 4 cores x ~2.15 GHz is easily stomping a 2nd gen i7 without any OC in multi-core loads. Not to mention none of the cores on my E5605 CPUs are going past 45C - and that's with stock heatsinks.
I actually use the BTX heatsinks to overclock BTX computers.
There was some success overclocking a T7500 with SetFSB. He had to use a PCIe HDD comtroller card to get the drives off of the SATA bus. X5687 is a resonable priced chip 4C/8T with good base clocks. TS won't work on dual CPU. It probably will get you turbo speed on all cores. https://www.overclock.net/forum/5-in...ion-t7500.html
Actually 4.3 Ghz may be realistic. It only crashes about 6 minutes into a 10 minute stress test when the temp starts to climb. I have a lot of headroom on the cooling side. I can add a lapped CPU, 20*F. better ambient, and crank up the fan speed ( alot!).
The TS overclocked T3500, W3680/90 seems to be the one that can keep up with newer systems. The 3 channel RAM seems to have a lot to do with it. https://www.userbenchmark.com/UserRun/18559176
X5687 is a resonable priced chip 4C/8T with good base clocks. TS won't work on dual CPU. It probably will get you turbo speed on all cores.
That's a very fast chip. But it's also quite toasty at 130 Watts. Two of those in my T7500, and I'll probably be roasting marshmallows even in the winter with the house windows open, lol. My original CPUs were 80W TDP each, and that's what I stuck to - but got two more cores (6-core CPUs) with the same TDP. The trade-off is of course low core clocks (only 2.5 GHz base, and I think up to 2.8 turbo.) But I'm really starting to like the concept of more slower cores over fewer faster cores now. Next mega PC project, if I ever take on it, will probably be a dual 16-core socket G34 setup. 12-16 core AMD CPUs' are dirt cheap on eBay right now (under $10 shipped.) The G34 mobos aren't, though... but still not that pricey either.
The TS overclocked T3500, W3680/90 seems to be the one that can keep up with newer systems. The 3 channel RAM seems to have a lot to do with it. https://www.userbenchmark.com/UserRun/18559176
Yup, Intel really made a mistake with how good of a platform they created with X55 chipset and 3-Ch DDR3.
Those old 6-core chips aren't as fast or efficient as newer chips, but probably still perform somewhere close to a modern 4C/8T chip... which is still worth something!
The aftermarket X58 MB are selling for $150 while Dell T3500 can be found for $80 whole systems. People have built 6C 12T 4.2Ghz systems for $130.
The key to success with Setfsb I think is getting control of the Voltage.
One method is to spoof the Vsense pin of the VRM controller with a pull down resistor (potentiometer) and install a separate Volt meter to disply the result. I want to try this on my XPS410 to get past Intels 1.6V. limit. I may or may not need to do that. I didn't do much exploring at 4Ghz 1.5875 Volts with a QX6800. I just got a CPUZ validation and called it quits. But with better cooling, and understanding that I may have been outrunning the PSU 18A. rail setup. I'm going to try again.
I thought The Delta 150x50mm fan was king of the hill. It turns out San Ace makes a 150x50mm 300cfm fan. I ordered a couple of them. 3900rpm,2.9A. 300cfm, PWM.
I can get the 2.9A. by using both MB fan headers on the mid tower Dells. The 3900 rpm won't be an issue because the 120x38 mm fan Dells came with are 3800 rpm. The Dell PWM profile is designed for that speed of fan. The XPS 410 I will be running the fan off of a Molex anyway. If enough is good, then too much is gooder!
[QUOTE=momaka;979839
That's a very fast chip. But it's also quite toasty at 130 Watts. Two of those in my T7500, and I'll probably be roasting marshmallows even in the winter with the house windows open, lol. My original CPUs were 80W TDP each, and that's what I stuck to - but got two more cores (6-core CPUs) with the same TDP. The trade-off is of course low core clocks (only 2.5 GHz base, and I think up to 2.8 turbo.) But I'm really starting to like the concept of more slower cores over fewer faster cores now. Next mega PC project, if I ever take on it, will probably be a dual 16-core socket G34 setup. 12-16 core AMD CPUs' are dirt cheap on eBay right now (under $10 shipped.) The G34 mobos aren't, though... but still not that pricey either.
[/QUOTE]
That's actually not a problem. Throttlestops primary purpose is underclocking laptops for battery life. On X58 systems you can limit TDP, TDC, and TPL to whatever you want, and limit the multiplier too. You can set up different profiles too. Anything from Turbo speed on all cores,down to any lower level you like. You can test it on the CPUs that you already have. We have one overclocker in India who is seasonal. During the summer he undervolts and underclocks his computer. I've split TS overclocking off from the laptop forums because it gets lost in the 1000 pages of underclocking data.It also gets lost in the BSEL overclocking forums. if you go to Notebook Review you will find many posts on doing this. You can have it both ways.
To put this in perspective the LGA775 45nm 4 core Q9550S 2.83GHz was a 65W CPU. The X5687 is a newer 4 core 32nm CPU so should be even more efficient.
BTX is just one fan. I got this because they're hard to find. The Delta 1.8A AFC1512DG 150x50mm is probably enough. But the Dell fan profile is kind of slow to speed up, so getting ahead of the curve may be useful for PWM operation. This size fan is actually a drop in fit for an Optiplex MT Micro BTX computer. I haven't done a full on O/C run with the big Delta yet. So IDK if there's any real benefit to the San Ace 150. But there's only one sure way to find out! I've stuck together an Opti 745 case with a Dimension E520 motherboard
that can handle a high wattage overclock. With a QX6800 and trying to go past 4GHz a little extra cooling couldn't hurt. With BTX I'm stuck with Dell OEM XPS heat sinks so the fan is the logical place to look for extra cooling.
Someone did a low buck Dell T3500 gaming rig.
W3680 Xeon 6C/12T 3.7GHz,12 GB DDR3, R9-380X 4GB GPU. Stock Dell PSU. https://www.techpowerup.com/forums/t...1#post-4409974
These are capable of much more, but for almost no money a pretty good effort.
Someone overclocked Dell T5810 workstation with an unlocked Xeon E5-1650v3 to 4.5Ghz.
Page 41 Post #1018 at the TechPowerUP Throttlestop overclocking thread. https://www.techpowerup.com/forums/t...235975/page-41
This user preferrs TS to Intel XTU for this. I wasn't sure which would be better.
over the weekend, i've had some time to play around with throttlestop on my dell latitude e6500 laptop. took me a few tries to get one right version that would work.
firstly, i saw on the techpowerup site downloads section that the latest version only supports win7 and newer, so i tried v8 and luckily when i ran it, it didnt say os not supported or an attached device is not functioning. typical telltale signs the os isnt supported but i got the error cpu isnt supported instead.
so i googled this error and it brought me to a few of retrorockit's posts on some forums who said that throttlestop is rebuilt with each version so support for older cpus gets removed and i saw one of retrorockit's posts to try v6 instead for core 2 cpus, so i decided to get that and it works and recognises the cpu in my dell laptop!
i then mucked around with it and saw that it doesnt display the cpu core temp properly. i saw that in the ini file, the tjmax was set to an incorrect value of 98 for core 2 penryn cpus so i changed that to 100 which is the correct value for core 2 penryns. i also rtfm and saw that u had to change the power scheme in power options to home/office desk in order for throttlestop to work so i did that. previously, i had it set to minimal power management as the laptop was always on ac wall power with the power brick, so no need for conservative power settings.
i then tried the lowest vid it could do at nominal frequnecy for my cpu which is a c2d p9700 @ 2.8 ghz. using occt cpu test as the stress testing program, i managed to drop it all the way down to 1.025v but the system self-rebooted after a short while! oops! so i tried one more step higher at 1.0375v but it still caused graphical glitches that hung and freeze the system when i tried to reboot to get rid of the glitches.
after restarting the system, i tried 1.05v and decided to leave it overnight to run. but when i got up to pee in the middle of the night, i saw it stopped after around 2.5 hrs with an error. so i then raised the voltage up one more step to 1.0625v and it ran for hours without errors after that. the core temps were still below 60°C, so it was good. refer to the attached picture. otherwise, previously at nominal vid, the cpu tcase temps would hit 72°C! and thats with the cpu blower fan in the laptop being forcefed 5v by hotwiring via the usb port! no way the laptop is going to last on lead free solder at that temperature!!
after that, i tried to see whats the lowest vid it could run on the lowest multiplier but alas, it wouldnt let me force the voltage to lower than 1v, so vid settings below 1v had no effect. but it was still a good test to see the temps and hence power dissipation in the lowest p-state. because that would give a good indicator of how long the battery would last. the temps at max load during occt cpu test were 45-ish °C so that was good... that was an approximately 12°C difference between the optimised voltage max p-state and lowest p-state, so that was a bit of power savings there by optimising the voltages. satisfied, i then saved the settings in throttlestop. will be sure to use it whenever i need to use the laptop on the go or when i need a brief power boost on the go.
I wonder how a 4+ GHz Core 2 Quad architecture compare to something like an i5-2500 or a similarly clocked 1st get i5/i7 with same core and thread count (so I guess no i5's then, lol ).
Yeah no SMT on Core2 machines...
Comparing my machines, seems 1MHz Core2 (FSB) =~ ⅔MHz 2nd/3rd gen Core (IMC) at the same frequency... Not sure about newer unfortunately but likely it will get down to ⅗ or maybe ½ MHz but not much more I think... but clocks and cores have gone up... and down...
There are 1st gen i7 unlocked Xeons to play with. The closest to C2Q would be the early W3570,W3580 Xeons. These are the same as the 45nm i7 X chips with HT, faster memory, and ECC support.The big advantages over C2X would be they're dirt cheap and have 3 channel memory support. Dell T3500 workstations can be had for $50 if you can avoid the shipping charges. TS can be text edited to only OC core 0,1 which with HT will still give you 4 fast cores to play with.
Almost as cheap is the 32nm 4c/8t X5687 which is not unlocked, but with 3.89GHz turbo speed on all cores with TS is not bad either.
Most popular are the unlocked 6c/12t W3680,W3690.
There were a couple unlocked early i5 CPUs if you already have the MB for that. One was a 2 core I think. They're mentioned in the TPU TSOC thread.
I almost forgot. Last and probably least is the inscrutable (China only) LGA775 E6500K.
That's right an unlocked 2.93GHz 266fsb 45nm 2 core with SSE3+ (huh?)
Son of X6800 I guess.
Comment