Thursday, December 22, 2016

How To Calculate an IMEI from a MEID/ESN (Luhn's Algorithm)

This is a short article on how to get an IMEI number from a MEID or ESN, which should be the same. For a lot of people, this can be found within the device. But there will be times when a device will just not show this information.

Why Do I Need This?
As stated above, most of the time an IMEI can be easily found in your device settings. Other times you may even be able to find this printed on the back of a device. However, there will be times that neither provide an IMEI.

As for what it is needed for, most of us would have to give it to a carrier to register our device on a network before we can use the network. There may be other reasons, but this would be the most common.

For Me
I recently had an issue with a brand new LG G5 (US992) device that comes factory unlocked, but is branded for US Cellular. There was nothing on the back, and the settings only showed an MEID and ESN.

Reading through some forum threads, some people stated that popping in a GSM SIM card would allow me to see an IMEI in the settings if the device was on a CDMA network previously. The LG G5 had been on US Cellular for about two days, so maybe this was a simple solution. But after putting in a GSM SIM card, nothing changed...

Luhn's Algorithm
Math. Math is the solution here. If you have either the MEID or ESN, you have the ability to calculate your IMEI.

An MEID or ESN will be 14 digits, while an IMEI will be 15 digits. The IMEI will have the same first 14 digits as an MEID or ESN, with a final 15th digit called a check digit. Luhn's algorithm is a way to obtain that 15th digit. It is a public domain algorithm commonly used today for numerous devices.

The Math
I will give you a simple visual showing of how it is done. We will first write out our MEID:


You will then double every second digit of the MEID:

1 2 3 4 5 6   7 8   9 1 2 3 4 5  
1 4 3 8 5 12 7 16 9 2 2 6 4 10

Add up all the digits as though they were all single numbers:

1  4  3  8  5  12     7 16     9  2  2  6  4  10         

Now add to the summed number until it is divisible by 10 and results in a whole number:


The number that we added to the 62 is our check digit:


Our IMEI would then be:


This is just an example, but shows exactly how one would use a real MEID/ESN to calculate the 15th digit (to get the IMEI).

Does It Work?
I have used this during two occasions. The first was for my LG G5, and I was able to validate it through an IMEI checker, which did display the correct LG G5 variant--and was later sold to a customer with that IMEI, and who has not complained since our transaction.

The second occasion was during a job that required activating phones at times. I had an elderly customer who did not have the IMEI but provided the ESN to me. Following the formula above, I calculated his 15th digit and was able to confirm that his device was fully activated. 

Cricket: How To Reach & Increase Data Speeds & Hotspot on Unlimited Plan!

I recently decided I wanted to get an unlimited data plan, not one with restrictions to where you drop down to 90's modem speeds, but one where you could really just use as much as you want without penalty. I had a few options, so with some time wasted, eventually landed with Cricket. Here is how it all went down...

The Competition
Initially, I had a few options: Boost Mobile, MetroPCS, & Cricket. My phone of choice: LG G5 RS988, the factory unlocked variant covering the most frequency bands available for the G5.

Boost Mobile is the same price as MetroPCS, but you could save $10 if you did autopay ($5) and used the "rewards" app ($5); which would require you to view at least one offer every day for 20 days out of your billing cycle.

The only issue with Boost Mobile (besides using the Sprint network), was that I would need an applicable device. I have an iPhone 4 that works on the Sprint network, but because that does not support 4G, it would be pointless. I have an international LG G3, but that does not have the bands Sprint requires--and there is a disclosure that Sprint has to confirm the device beforehand. So, Boost Mobile was not a viable option for me.

MetroPCS was the next logical choice at $60. There was no credit for autopay, but since it used T-Mobile's network, it would work with my devices, which are all unlocked. I decided to try my luck and visit my closest store that is only 5 minutes by car.

I liked that even with the activation fee there were no taxes or surcharges added ($80). We tested the data at the store, and made a call (that took 20 seconds before getting a ring from the receiver). Once I left the store, I did an immediate speed check. It was pretty good at 36+Mbps, and higher upload speeds than my home Internet.

But as soon as I got home, the speeds had dropped to just above 1Mbps, and less for upload. That was not a problem since I did have home Internet, but what was a problem was calling or receiving calls. I could not do either after having tested with a friend. Every time I tried to make a call I would either lose the 3-4 bars of signal and 4G/LTE icon, getting an error message that "Mobile network is unavailable." Or, it would never ring, then end the call.

Of course I called tech support to diagnose the issue. They checked all my settings, had me restart, and verified my UCCID (SIM) number. Doing what they called a "network refresh", they told me to try to make a call in 30 minutes to see if the problem persisted. It did, so I called back and reiterated the issue and what had been done. In the end, after being asked if my device was 4G-capable, I had to cancel my service that was just over 3 hours old. (I am still fighting to get my money back!)

I was now left with Cricket as my last option. The good news is the Cricket is truly unlimited. Unlike Boost Mobile, where they would restrict your speeds after 23GB of usage, or MetroPCS, where they restrict you after 28GB of usage, Cricket does not have a limit. The only restrictions they have are speed limits (8Mpbs for 4G devices, 4Mbps for non-4G devices), which applies to all their plans. This was okay with me as I would not truly need higher speeds unless I was multitasking (e.g. watching FHD video and doing multiple app updates simultaneously).

The only other problem I could foresee was that in my area, it may have poor speeds in general, disallowing me from utilizing the data as I saw fit. A Tom's Guide article indicated that Cricket was at the bottom of the list in terms of speed, not even averaging its maximum speed limit! AT&T (the network Cricket uses), did pretty well, but its subsidiary was quite paltry in comparison. This was for a metropolitan city about an hour's drive away, so it could better or, likely, it could be worse.

There was no other available option for unlimited data, so I had to take my chances. I knew my phone would at least work as I had tried it with a AT&T prepaid SIM card previously. Again, I went to the store and signed up for the service. Since there was still a $5 credit on my AT&T prepaid account, I asked if there was a way to transfer that over. They were unable to do so, but did knock off the $10 SIM card fee they charge. This left me a charge for a little over $95 after taxes ($70 for the plan, $25 for the activation fee). Which may mean that they will charge taxes on the next bill, as AT&T prepaid does, but I was given assurance that they notated my account to waive any SIM fees if I ever needed a new one, as well as some additional fees if I had to add any new lines. And with autopay, I will save another $5, making it only slightly more expensive that what I would have been paying with MetroPCS.

On a side note, besides Cricket charging more than the competition, I would also lose the ability to have HD Voice that I could retain if I had Boost Mobile or MetroPCS. This is not a concern for me, but it does change my signal to "H" during my calls (and probably to "3G" in certain locations).

We had tested the data and made sure I could make calls (I already had some texts from Cricket about signing up). When I got into my car I did a speed test, which was slow at around 3Mbps for download--and somehow faster for upload--but it would get the job done. The ultimate test was when I arrived home. I first did a speed test, which was still about the same, which also meant better than MetroPCS. I then made a call to a friend, and it worked beautifully!

The last things I did included checking my settings for my network. I was originally on LTE/GSM/UMTS, but I found changing to Global--the default for factory unlocked phones--did a bit better in terms of speeds. I could also switch between these two choices after a call to get my 4G LTE signal back instantly.

And I also found that while the unlimited plan was optimized for 480p video, moving up to a 720p YouTube video seemed to work without the need to buffer constantly. Trying 1080p, buffering happened quick and often. Better speeds may correct that, but I can survive with HD on a phone or tablet.

How To Reach Cricket's "Fastest" Data Speeds
I mentioned earlier how I was getting fairly slow speeds from Cricket when I tested at the store and at my home. Even if I could only get 8Mbps downloads, I would prefer to get that as much as possible. So I set out to see if I could...

There were two free apps that I tested, and both worked. I believe they flush some settings and give you a clean slate with the network. From there, I was able to achieve much better speeds with great side effects.

The first app is Network Signal Speed Booster. It is a 16MB+ app with ads, but it takes a single button press to get it started. An ad will come up before it is done, then you can leave (not exit) the app and use your data to its fullest. I tested with's app--there are other good apps, but this is a popular one--just ensure to enable your GPS (or location as it will likely be called) to get more accurate results. With this app I was able to get over 8Mbps download speeds, with upload speeds at 4+Mbps.

The second app is Internet Speed Booster Free. It is also is a 16+MB app with ads, and also takes a single button press to get started. The app aesthetic is a bit more polished, but I feel like its ads take more time to exit, that being a second or two longer. On top of this, it does show in the corner of the screen when activated. I do like this feature of the app as I like knowing that it is still running. (When I returned to the other app mentioned at one point, I had to restart it, even though my testing showed it seemed to still be running.) With this app I was able to get similar results in terms of speeds.

The major differences between the two apps was that the first app seemed to have slightly lower latency times when tested; however, the second app seemed to have slightly better download and upload speeds.

With either app, I was then able to run FHD (1080p) videos without buffering. I expected this would be true if I could reach Cricket's speed caps, but only now was able to confirm it. I can also safely say that with either app I was able to achieve higher Mbps speeds than the speed cap, nearly hitting a constant ~9Mbps download and 5+Mbps upload with the second app during two tests!

I should state that you may get different speeds in different locations and have to retest when moving from one spot to the other. Traveling and trying to get the most speeds out of where you are may also prove fruitless since you will likely end up with different speeds when you finally reach your destination.

How To Surpass Cricket's Fastest Data Speeds
I had been extremely pleased with my findings, and did not plan to push it any further. Yet, when I was watching a YouTube video I noticed that it was available in 1440p (2.5K). Of course when I tried it, there was constant buffering in between seconds of video playback. While I did not expect to be able to watch 1440p without buffering, I thought it would be fun to see if I could.

I brought up an article to find out what the average speeds needed to be in order to watch 1440p. It showed 1080p as needing 8Mbps, which rang true from my research, and that 1440p needed 10Mbps. Could I really get above Cricket's speed cap and get smooth video playback?

I discussed the possibility with a friend, and we both landed on the only plausible solution being different APN settings. I had already tried AT&T to no avail, but maybe there were some out there for Cricket I had not yet tried.

I went through a lot of articles that either gave me the same exact APN settings Cricket's SIM has me use, or just rubbish hacks that were of no use to me. Finally, after reading through a thread further than necessary, someone gave some APN settings that they were using. The APN settings I have were tried by this person and did not work for them, but they have found some that did.

There were two APN settings, one for just the Internet, and another for just the MMS. The Internet APN settings added without complication to my phone, but the MMS would not stick. It is rare that I send a MMS, and when I do send a picture, it is normally through a messaging app. 

After plugging in the new Internet APN settings and selecting it, I made a test call and SMS, both of which worked successfully (I did not try a MMS). I then ensured that one of the two apps I listed above (to reach Cricket's speeds caps) were in effect before trying a speed test.

The results were amazing! I was able to hit anywhere from 9+Mbps-11+Mbps for download speeds! What was odd during my tests was that it seemed that choosing the Global option gave better upload speeds, but not always better download speeds when using LTE/GSM/UMTS--albeit a small difference, if at all.

My next test was to go back to the YouTube video and try 1440p video playback. With everything in effect, I was able to watch 1440p without buffering once! The new APN settings was working, and working to help go above the speed caps imposed by Cricket.

The following are the Internet APN settings that I added to not only increase my Cricket speeds, but surpass the Cricket speeds caps (I have excluded any settings to be left blank):

  • Name: AioInternet
  • APN: ndo
  • MCC: 310
  • MNC: 150
  • Authentication type: PAP
  • APN type: default,supl,hipri,fota

If you do end up needing the MMS APN settings as well, remember to add them but still select the Internet settings as your primary option:

  • Name: AioMMS
  • APN: ndo
  • MMSC:
  • Multimedia Message Proxy:
  • Multimedia Message Port: 80
  • MCC: 310
  • MNC: 150
  • Authentication type: PAP
  • APN type: default,supl,mms

Note: It is possible to change "ndo" to a name of your choosing without affecting the other settings. This can be helpful if wanting to quickly know which settings you added and which settings came from the SIM card without first looking into the settings of each.

You may notice that combining these two APN settings gives you the normal Cricket APN settings. So, for whatever reason, using the Internet APN separately may somehow help increase your speed limits. This leads me to believe that somehow the MMS settings helps restrict or burden the data bandwidth, but I will never be sure.

How To Hotspot on Cricket's Unlimited Plan
If you check the fine print for the unlimited data plan through Cricket, you will read that the hotspot function is not supported. For their $50 or $60 plans, you can have the ability to hotspot for an additional $10/mo. And if you dig, you will find that you can only use up to 8GB for the hotspot feature.

I wanted to see if I could find a way around this, and I did! At first, I tried Bluetooth Internet sharing, but I could not make it work at all. But if I used my normal hotspot feature, I was able to connect to that hotspot from another device. I checked my account through the Cricket app, and have not yet seen any extra charges for a hotspot.

Maybe this is because the Cricket system is not setup to add that charge to unlimited plan..? I am also unsure if this also has to do with the phone I am using, since it is not from a carrier, and is completely unlocked (less the bootloader). Although, I would debate that the former is a more likely cause.

I also did try using a hotspot app before using my own hotspot, and found that FoxFi did work without issue as well. So, if by chance, the phone is the reason I can get around the hotspot problem, then using an app like FoxFi might help out others.

The Final Chirp
While I have not used Cricket for very long, and I will see if I keep them for more than a month, depending on their coverage in different locations, I am so far happy with my choice. In addition, Cricket does have a rewards app that may give me some benefits in terms of costs I have yet to weigh.

Friday, November 11, 2016

TIME SENSITIVE: You Have 8 DAYS To Put One Over on Toys 'R Us!

It has been a good while since I have had a chance to write an article, and I thought I would start with something that ends on 11/19/2016! As the title suggests, it has to do with Toy 'R Us. This will not be the most amazing secret, but it will be nice for those doing some early holiday shopping and wanting to squeeze every penny possible!

The Coupon
My daughter recently received a $100 gift card to use at Toys 'R Us. We finally got a chance to go and of course she wanted everything she saw.

The in-store magazine I picked up at the front had two coupons on the back: One had expired and required a $100 purchase for a a $25 gift card, and a second that requires a $75 purchase for a $10 gift card. The latter still valid for another 8 days!

After getting a large amount of toys (more than what the gift card would afford), we decided to go ahead and call it a day. Armed with a gift card and coupon, we proceeded to get rung up and head home.

The Correction
I had to drop off my family, then my sister, then head back to my place. However, while examining the receipt, one toy seemed to be overpriced. My wife and sister verified that the price was right, but I still did not think so.

When I got to my sister's place, I checked the price of the toy on Toys 'R Us' website. It was indeed cheaper than what it was at the store. I called the local Toys 'R Us to confirm I could get the price difference, and was told I could.

I had to go back-and-forth because I needed to grab the toy--just in case--and I had left the receipt at my sister's. When I finally got everything, I headed back to Toys 'R Us and had no trouble getting my credit (which was placed on a gift card).

The Mistake
Before I left Toys 'R Us, I asked why the price difference placed on the gift card was smaller than what basic math would have me believe. I was shown a price below the toys that was less than their actual retail prices, and was told it was based off of that. Supposedly, I was (told I was) given some discount that I was not supposed to have...

Not being as dumb as most, I decided to get home and do the math. I added up the entire retail amounts, which did give me the total I was charged. Immediately I knew something was wrong, and it seemed that I was still owed more than what I was given...

I called the local Toys 'R Us to explain my findings, and was about to talk to the person who I had dealt with previously, before realizing something... I finished up the conversation quickly by playing dumb and having her "math" reexplained to me while feigning tiredness.

The Discount
For whatever reason, the current coupon is processed incorrectly by their POS systems and provides you a total WITHOUT SALES TAX! Apparently, however much you buy, the total amount of the retail prices will be the final amount that you pay!

The coupon is for a $10 gift card, and the way the system needs to allow this is by pretending that you are purchasing that $10 gift card. So, the system decreases the prices of all the applicable toys until it the total price is $10 less than what it should be. It can then add the $10 gift card to the total price as if you were paying for it.

Now, the error occurs because the equation should equal the total retail price before applying the sales tax. Yet, the equation applies the sales tax in order to reach the total retail price!

This is the reason I did not mention anything to the lady on the phone during our last conversation. It struck me that the extra $3+ I am owed is much less than what I owed Toys 'R Us in terms of sales tax.

The Check
Besides having all the items and taxes outlined on the receipt, I went ahead and checked a couple of items on their site that I had bought in their store. The prices for these toys were correct, and when I got towards the end of checking out online, there was indeed tax applied to the toys!

Get It In Now!
I am not sure if the gift card or toys that were on sale had any impact, but I do know the coupon is the key here. It is unlikely that any other Toys 'R Us will deviate from using the same method, so take advantage while you can.

As a side note, if you live in an area that uses a high tax rate, this will save you more than those living in lower tax brackets.

Go get those toys!

Saturday, July 30, 2016

How to Make a MULTI-UEFI Bootable USB! (Windows/Easy Method)

Making a bootable USB has never been easier with the amount of software freely available on the Internet. My favorite happens to be Rufus, but there are many to choose from. And while there are several multi-boot USB programs for MBR, or single-boot GPT USB programs, making a multi-boot GPT USB in Windows is somewhat unheard of. I want to share with you how to make a multi-boot GPT USB easily in Windows.

In the past there was only BIOS, but in recent times UEFI and EFI have become the new standard for motherboards to use. The main reason for this is that BIOS has space constraints that UEFI improves upon. Much like many protocols, BIOS was made and meant for a certain era, while UEFI attempts to modernize and supplant BIOS.

The simplest way to think of UEFI is like a better BIOS.

Multi-UEFI (GPT) Bootable USB
I want to start off stating that Linux and Unix users have had the capability to use a multi-UEFI bootable USB's for quite some time, albeit through a CLI (Command Line Interface). But Windows users have really not had the ability to do this--until now...

Why is a multi-UEFI bootable USB a good thing? Because you may have several rescue CD's or Windows/Linux Live CD's that you want on one USB as opposed to several. For example, if you were IT for a company that fixed systems, it would be a big hassle to use several USB's, and even worse if none were labeled!

Is It Necessary To Use UEFI?
I would argue that currently it is unnecessary. I have never come across a motherboard for either a desktop or laptop that primarily used UEFI (or EFI), but did not have a "legacy" mode that would allow for a BIOS-compatibility mode.

In essence, this is a switch to change between BIOS or UEFI.

That being said, it should be noted that it is likely that this will not always be the case. It may take a long period, but I doubt that this legacy mode will always be around. And, for the sake of simplicity, it is much nicer to be able to just plug in a USB and not fiddle around with UEFI settings beforehand to get a USB to be recognized.

While having to switch a couple of settings does not sound like a chore, it certainly can be. Last week I had to do a lot of testing in legacy mode; so I had to go into the UEFI, switch to legacy mode, test a USB, go back into the UEFI, switch back to UEFI mode, then restart to get back to Windows. It was not fun.

Introducing YUMI
Some of you may already know of YUMI, it is most commonly known as a multi-boot USB for BIOS (MBR) systems. It is simple to use, portable, and has an intuitive interface.

What most people may not know is that YUMI now offers a UEFI beta version of their program. This only works with 64-bit systems, so be aware that any 32-bit system will not be able to utilize this version of YUMI. (A 32-bit version is being worked on). You must also disable Secure Boot if it is enabled in your UEFI.

There are a few (known) restricted rescue CD's and Live CD's that will not work with YUMI UEFI, which is not something the YUMI creators can fix.* That is because these CD's were not made to work with UEFI, so support for this CD's would have to come from the CD creators.

And while there is a list in YUMI of CD's that can be used, CD's that are not on the list are not guaranteed to work or not work. You will have to try such CD's to find out if they can boot and work properly.

*In this list, CD's that cannot work in UEFI will have a note next to the name stating "BIOS ONLY".

Using YUMI UEFI is very simple, but I will outline the steps for those who would rather get a sense of what to do first:

Go to the YUMI homepage.
Scroll down to "Beta Download -> UEFI YUMI - Windows".
Download YUMI UEFI by clicking on the "UEFI YUMI - Windows" text only.
Double-click the "UEFI-YUMI-BETA-0.0.0.#.exe".
Press "I agree".
Click the dropdown arrow under "Step 1".
Select the USB flash drive you have inserted.*
Click the dropdown arrow under "Step 2". This will show you a menu of every known rescue and Live CD that YUMI supports.
If you do not have one already, download a rescue or Live CD you wish to use.
Click the dropdown arrow under "Step 2" again.
Select the rescue or Live CD you have ready. If your rescue or Live CD is not in the supported list, you can either choose a supported rescue or Live CD that from the list that may be similar so you can move onto "Step 3".
Press "Browse" under "Show All ISOs?". Check off "Show All ISOs?" if your rescue or Live CD is unsupported, or if you have changed the original name.
Browse for your ISO and double-click it.
Check off "Format #: Drive (Erase Content?)".
Click "Create".
YUMI will not show a progress bar, and when full, it will display a "Finished" button. Click "Finished".

*Note I: If you do not have a USB flash drive inserted, do so now, or check off "Show All Drives?", then select any drive. But be careful, you do not want to accidentally format your OS drive.

Note II: An external USB HDD/SSD (and in fact an installed HDD/SSD) can also be used to create a Multi-UEFI bootable device. To find these options, check off "Show All Drives?" in YUMI.

To test your multi-UEFI bootable USB, first ensure that Secure Boot is disabled in your UEFI. Reboot your computer and click the F# key to get to your Boot Priority screen (or go into your UEFI and either rearrange for USB devices to be found first, or select the USB to boot from it). If the multi-UEFI USB is working, the Boot Priority screen will just show a USB, instead of a designated name that you might find using a BIOS.

While in theory this should work with any USB and USB port, I would recommend using a USB 2.0 flash drive, however, a USB 3.0 flash drive may work fine. But even moreso, be sure to test on a USB 2.0 port, if possible.

The reason I mention this is because I have found USB 3.0 ports to be finicky at times, especially during POST. I did not have this problem with my multi-UEFI bootable USB, but I tested with a USB 2.0 flash drive in a USB 2.0 port...

Too Simple?
No, it really is that simple. There are a few useful rescue CD's that cannot be booted from UEFI, but the majority you may know of are all available for UEFI.

For someone like me who may need to diagnose or repair a UEFI system every-so-often, a multi-UEFI is worth its weight in gold! And I know that more than a few people have been waiting for an easy method to make a multi-UEFI from Windows.

Friday, July 29, 2016

CPU Lapping vs Delidding: One, The Other, or Both?

CPU lapping and CPU delidding are two things that can help with your CPU performance in terms of thermals, and thus, overclocking ability. The majority of computer users will have no need to perform either, but those of us who strive to squeeze every drop of performance out of our CPU's may want to try one, the other, or both. Today I want to give an explanation explain on what they are and what they can do.

CPU Lapping
Simply put, CPU lapping is the process of taking a CPU and making it perfectly flat. You may be thinking, "Isn't it already flat?". For the most part, it is. However, there are often times where it is not perfectly flat, and because of that, it can have an affect on your CPU performance.

First off, a CPU will be connected to some sort of heatsink in order to properly dissipate the heat it creates. If the surfaces of the CPU and the heatsink are perfectly flat, then the contact between the two will allow for the best heat transference possible. Yet, if the CPU, heatsink, or both, have surfaces that are not perfectly flat, then they may only be making some contact, allowing the CPU to retain a portion of the heat it is meant to get rid of through the heatsink. Generally, more heat to the CPU equals less performance.

Second, before one would perform CPU lapping, they would want to check if their surface is flat or not. An extremely easy way to do this is to take a metal ruler and check different areas on the CPU to see if it is evenly flat. If it is, then CPU lapping is unnecessary. The same can be done to check a heatsink.

The actual act of CPU lapping is relatively straightforward: Take sandpaper (you can use one type of grit or have differing grits for smoother results), place the sandpaper on a smooth surface like a table (taping it is a good idea), then proceed to move the CPU up-and-down on its surface over the sandpaper until the CPU has a reflective surface.

CPU lapping is meant to help bring down temperatures, and it can, but the amount it can depends on the severity of the CPU surface. From what I have read, it seems that most people get very little gains compared to the time and energy used to achieve those gains. I would say, on average, most people get around 1-2 degrees less on their temps.

CPU Delidding
Delidding a CPU is defined as taking off the lid--otherwise known as the IHS (Integrated Heat Spreader)--and replacing the thermal paste inside with a better grade of thermal paste.

This was not something that was always done as it was not until the Ivy Bridge generation of Intel CPU's that made this (somewhat) necessary. Prior to that generation, all "lids" were soldered on. For some reason, Intel decided to change this approach and use cheap thermal paste, causing worse heat transference.

The process of delidding can be done in a few ways. The most common is by use of a razer to carefully pry open the lid. More recently, a few companies have decided to produce and sell delidding tools that make it foolproof to perform this process: Rockit 88, der8auer, & Enter Setup.

I happened to pick up the Rockit 88 version, as it was the most inexpensive and the only one located in the USA. The der8auer version I believe was the first on the market, and is from Europe, but at least twice as much as the other options. The Enter Setup version is probably the cheapest choice if you happen to live in or near Europe.

On a side note, if you do happen to get the Rockit 88 delid tool, you have the option to also pick up the "re-lid" tool, which is meant to help to place the lid back on. I recommend not buying this additional tool as it should make little difference if you plan on putting the CPU immediately into a motherboard after delidding... 

If you have a 3D printer and want to make your own delid tool, here is a project that people have had success with: 3D Printer Delid Tool Project

Once the lid has been popped off, you will clean the small rectangular surface area that has the current thermal paste (which is probably solidified). Apply some new thermal paste and place the lid back on. The tricky part is that you will want to apply pressure to the lid so that it is completely flat. Most people who perform this procedure are ready to place the CPU into a motherboard and attach a heatsink, which helps keep it flat.

Again, from my readings, most people seem to notice a 5-10 degree drop in the average temperatures. This is a decent drop, and replacing factory thermal paste will often produce such results.

Is CPU Lapping/Delidding Important?
I would say that most people will have no need to really do either procedure. Even horrible temperatures would suggest issues elsewhere that need investigating. But it will matter to enthusiasts and overclockers. People who have the desire for the lowest temperatures at all times!

An enthusiast may just want this for bragging rights, but an overclocker will want to try these methods since lower temps can help achieve higher CPU overclocks. There is obviously no guarantee, and a stable overclock will largely be based on the binning of the CPU; but I can attest that lower temperatures definitely help to push a CPU further.

I guess if I was looking for the best temperatures possible, then a CPU lapping and delidding would be in order.

But I am a practical guy, and I actually just faced this situation recently on a new desktop with an i5 3570K. The CPU lapping seems easy enough to do, but for getting 1 degree less on my average temperatures, it is just a waste of my time. I have also heard that it can depreciate the resale value of a CPU compared to one that has not been lapped.

Instead, I opted to only do a delidding. I used the Rockit 88 delid tool as I did not want to harm anything with a razor. I then applied some thermal paste and placed the lid back on the CPU.

Was It Worth It?
I did end up buying the re-lid tool for Rockit 88, so I let my CPU sit overnight before inserting it into an ASUS P8Z77-V motherboard. My goal was to see if I could overclock the 3570K to at least 4.8GHz, so I attached a Thermaltake Water 2.0 Extreme AIO closed-loop solution. With the Thermaltake AIO running in silent mode, my temps stay in the low 30's.

While I soon realized my 3570K was an average overclocker, the combination of the delidding and the Thermaltake AIO allowed me to achieve a stable 4.9GHz overclock! For those of you familiar with the "legendary" i5 2500K, most people would claim a 2500K running at 5GHz could match a 3570K running at 4.8GHz. When using the benchmark in CPU-Z I was able to outperform an i7 4970K, and come close to reaching an i7 6700K.

Not bad for a CPU from 2012 that has no Hyper-Threading.

My foray into delidding was a successful one. It cost a bit of money due to the reason of not just using a razor (of which I have none on hand). And I completely avoided CPU lapping as it really seems to do so little. I mean, if the CPU is not relatively smooth to begin with, then that CPU needs to be returned to the manufacturer!

If you decide to try either method, I would advise watching some of the YouTube videos to get a better feel of what you need to. And if you do not try any of the techniques, rest assured that you are not suffering any great performance losses (unless you are an overclocker!).

Thursday, June 30, 2016

M14x R2: Overclock Locked CPU & 16GB RAM @ 2133MHz

The Alienware M14x R2 laptop was (and still is) a great laptop. It had some good options and was great for either gaming or daily work activities. But some things that most people were (and are) unaware of about the laptop, which can make it great, have never seemed to be truly revealed. While this laptop is getting older, it can still provide a lot of power, if you know how to unlock it...

M14x R2
The Alienware M14x R1 came in with an Intel Sandy Bridge CPU with a NVIDIA GT 555M graphics card. Later on, the M14x R2 came with an Intel Ivy Bridge CPU with a NVIDIA GT 650M graphics card. The biggest upgrade obviously being a GPU (as the M14x R3 was said to have a weaker GPU than the R2), and the move to Ivy Bridge allowed for the jump to PCI Express 3.0.

The screen resolution was either 1366x768 (720p+) or 1600x900 (900p), which was 14" in size (hence the name M14x). It normally came with 6GB DDR3L RAM with a voltage of 1.35v, but it can take 1.5v.

My M14x R2
I did not get my M14x R2 until years after it had been discontinued. My father had seen my M17x R2 and decided he wanted an Alienware laptop. He had me purchase one on his behalf. Once he finally received it, he tried it for a few days before moving back to his older laptop. Recently, he gave it to me, and it is practically in new condition!

The specs on mine are almost as good as they could get when bought. It came with a i7 3840QM, a GT 650M 2GB GDDR5, and a 900p screen. There was a HDD instead of SSD, but I have since installed a M.SATA Samsung 850 EVO 500GB SSD. It also had 6GB (2GB+4GB) 1600MHz RAM, which I changed to 16GB 2133MHz RAM.

The only issue that this laptop has is the trackpad has you accidentally moving the cursor around at times because your hand will touch it while typing. It had another mouse issue, which was the main issue my father had with it, but I was able to disable this in the Windows settings.

As for the graphics card, there were supposedly different variants. There was a GT 650 1GB GDDR5 and a GT 650 2GB DDR3 version available. Thankfully, when my was bought, it had the best version, which was th GT 650 2GB GDDR5 variant. While that sounds great, it is not a powerful graphics card. It can be pushed to do better (with an "unlocked" BIOS), but it falls short of the options of today. And even though the CPU is upgradeable, the GPU is not. It is soldered in. However, it can still play most modern games on medium or low settings.

Possible to OC Locked CPU?
The 3840QM is a great i7 Ivy Bridge CPU. It has a stock clock of 2.8GHz with a turbo up to 3.8GHz. In comparison, the unlocked 3920XM has a stock clock of 2.9GHz with a turbo up to 3.8GHz, and the unlocked 3940XM has a stock clock of 3.0GHz with a turbo up to 3.9GHz.

Essentially, my 3840QM will run as fast as a 3920XM (in terms of turbo) when it needs to, and almost matches the 3940XM. If I were to replace the 3840QM with either Extreme Edition CPU (XM) it would still cost me at least $300, if not more. That is far too much for the little gains I will get from a XM CPU. I may be able to push an XM to around 4.2GHz safely, but it still does not seem like a great idea economically.

While I was researching about my CPU recently, I came across a couple articles about overclocking the 3840QM. One gave very little information and was just someone asking if it was possible, with a response that it was not. But the other had very useful information. Someone stated and showed that they were able to overclock their 3840QM up to 4.0GHz! That is impressive just for the speed, but even more impressive knowing that this CPU has a locked multiplier and should not have any way to do so!

So, I decided to try this for myself and see what I could. I did several tests and watched my clocks for differences. In the end, yes, I was able to push my 3840QM to a turbo of just over 4.0GHz!!! (3.99GHz consistent!)

I have since read that the Ivy Bridge era did allow for these overclocks, but no generation after has been allowed to do so.

Overclocking Locked CPU
This will require you to download ThrottleStop. ThrottleStop 6.00 is the last stable version released, however, if you search the TechInferno thread on ThrottleStop, you can find an active ThrottleStop 8.00 Beta version. It is unnecessary to use the newest beta version, as it does not add anything that we will need.

The first thing to be done will be to enter the BIOS. Once inside, find the option for turbo and disable it. The next steps all relate Throttlestop:
  1. Open Throttlestop.
  2. Check "Clock Modulation".
  3. Check "Chipset Clock Modulation".
  4. Check "Set Multiplier".
  5. Check "C1E".
  6. Uncheck all other boxes.
  7. Under "Set Multiplier", ensure that it is at the highest multiplier setting, or has a "T" next to the number.
  8. Click "TPL".
  9. Change the cores to all run at "42".
  10. Click "OK".
  11. Click "Turn On".
Note: For whatever reason, I found that "42" was the multiplier number required to hit 3.99GHz consistently. Anything higher would do the same.

If you want to test the validity of the frequency achieved, try running a benchmark like Intel's Extreme Tuning Utility (IXTU) that tells you the maximum speed it hits. You can also use ROG Benchmark that will show its own CPU stat, or keep Throttlestop open and watch what frequency the CPU hits.

Running 2133MHz RAM!
While Ivy Bridge does support 2133MHz, that does not mean it is always achievable for other laptops. Factors can be the RAM itself, the motherboard, etc. Just because a CPU can do something does not mean it will. For example, Ivy Bridge can run up to 32GB of RAM, but that is impossible for an Alienware M14x R2 to achieve due to its motherboard.

Many forums discussing this very laptop claimed it was not possible to run RAM at 2133MHz, but I know this to be untrue since I have done so. I will explain how I did this so others can emulate it.

To do this, I bought a pair of 8GB Kingston HyperX Impact Black DDR3L RAM @2133MHz. If installed immediately, you may not be able to post. Install one stick and set the RAM settings in the BIOS to run at 1866MHz. This will allow you to post while using both (or one) sticks of RAM:
  1. Download IXTU.
  2. Open IXTU.
  3. Change to the "Advanced Tuning > All Controls" tab.
  4. Scroll until you see settings for Memory, DIMM or RAM controls.*
  5. If you hover over each timing setting of the memory, you will see the recommended values for that timing setting. Record the recommended values.
  6. Change to manual/custom timing settings for the RAM.
  7. Input all recommended values of the memory timings recorded in step 5.
  8. Save the timing settings.
  9. IXTU should prompt you with a reboot, select "Yes".
  10. Upon reboot, the laptop may not fully boot and get stuck. If it does, give it a minute before doing a hard shutdown (press and hold the power button).
  11. When you boot back up into Windows, open IXTU again.
  12. The recommended values should now be inputted, but you may have a value(s) that is not sticking. If not, you are done. If so, continue on...
  13. Note the setting(s) that is not sticking and turn off the laptop.
  14. Boot into the BIOS.
  15. Go to the RAM timing settings.
  16. Change the timing setting(s).
  17. Save the BIOS.
  18. Reboot. If the laptop does not boot up fully, wait a minute before shutting it down.
  19. Reboot the laptop.
  20. When you boot back up into Windows, open IXTU again.
  21. All the recommended values should now be showing. If not, repeat steps 14-20  again.
*If you do not have any memory options, then your laptop does not support this. This should only be true if you are using this tutorial for a laptop other than an Alienware M14x R2.

This is the process I went through in order to get 2133MHz RAM to be recognized and work properly. It will work, and enable the fastest RAM available for Ivy Bridge. But if you do not check the timings, you will either get the RAM speed pulled down to something slower, or if you have not changed all the timings to the proper values, you could cause instability.

Hard Shutdown
These are a couple of secrets people are unaware of. It is very possible to run locked Ivy Bridge processors at higher speeds than their turbo, and you can get the best RAM available for this generation!

To be able to do this, and on a laptop no less, is an amazing feat. With a setup like this, what would be the point of upgrading when this is just as powerful now as it was when it was first introduced?

Wednesday, June 29, 2016

How I Got a GTX 980 Ti (NIB) for $350!

The last time I wrote about getting a deal was for a GTX 970, and a GTX 980 the time before that. And today I will complete the high-end trio with a GTX 980 Ti! At the time of this posting, this is still a great deal, just as it was when I got it.

Why did I need a GTX 980 Ti? Well, because a couple of things occurred that led me to having a reason to need one...

I had initially bought my GTX 970 because I was doing an eGPU (external GPU) setup, which I had planned to post about. I successfully used this setup with an ExpressCard interface on my Alienware M17x R2, and was attempting to get it to work via mPCIe (mini PCI express) on my Alienware M14x R2. After several issues, one rendering my M14x R2 useless, I decided the best thing to do was to cut my losses and go for a different setup.

The setup I wanted to do was to get a new Alienware 17 R3 (AW17R3) laptop with an unlocked CPU, for future-proofing, and which would have the ability to use an Alienware Graphics Amplifier (AGA). An AGA essentially lets you connect a desktop graphics card of your choosing to use instead of the soldered in GPU the AW17R3 offers. With an unlocked CPU that could be overclocked to match the performance of CPUs released later on, and an AGA that would allow me to use any future GPU, it would be an ideal situation to run for at least 2-5 years before ever considering another "upgrade", if desired.

Since this was my goal, I needed money to fund this objective. So, I decided to go ahead and sell off my laptops, the eGPU, as well as some other things. I was well on my way and even went ahead and purchased an AW17R3 so that I would not be without a laptop. But another problem occurred that changed things a bit...

At that time I had a three-monitor setup that would be the equivalent of 1400p+, something the GTX 970 handled quite well. However, one of my monitors went a bit nuts, and I decided to go ahead and grab a UHD TV, often mistakenly called a 4K TV by most people. The TV was refurbished and cost about $230+ total! Another great deal! Oddly enough, after I made the purchase, the monitor sorted itself out and now works, but I had no intention on ditching my UHD TV.

But keeping the UHD TV presented another problem. If I were to use it for gaming, I would need a stronger card than the GTX 970. I love the GTX 970 and think it gives great performance, but it would just not be able to provide decent framerates at UHD resolution (at least, not at the highest quality settings). Having tested the GTX 980 at UHD, I knew that it would not be enough to give a good experience either. The minimum I could use was a GTX 980 Ti.

The Hunt
I sold the GTX 970 about a day or two before the official GTX 1070 launch, which also meant the GTX 1080 was already out. I paid just under $240 total for the Zotac GTX 970, and I sold it for $210. I immediately started searching for a GTX 980 Ti since people would be trying to get rid of them in order to upgrade to a GTX 1070 or GTX 1080.

I ended up messaging about 15 people or so, getting different responses to my offers. Finally, I got a reply that accepted my offer of $350! It was for an EVGA GTX 980 Ti SC, a reference edition. This was perfect, the AGA does not seem to have any type of cooling system along with its built-in PSU, so I would rather have a blower fan that will move out the air, rather than an aftermarket fan(s) that will move the air around in the enclosure.

The best part of this deal? The graphics card was never used! It was only opened to ensure everything was included and put back because the owner did not have the time to put it into his rig that he was building. When the time came to get a card, he went ahead and bought a GTX 1080 instead and decided to sell the GTX 980 Ti.

I will also add that because it is an EVGA graphics card, the warranty can be transferred. But this card had never even been registered, so I am the official first owner!

GTX 980 Ti vs. GTX 1070
A GTX 1080 would be great for UHD resolution, but they continue to command a $700+ price tag. So, the real debate would be why get a GTX 980 Ti over a GTX 1070? The GTX 1070 is newer, should perform better, and will do it at less watts. This is all true, but they also cost at least $450, and this is for a preorder that you may not get for weeks.

Even more important, I need a graphics card purely because I need it for a UHD TV. If you compare the GTX 980 Ti and the GTX 1070 in gaming benchmarks, they are fairly close. Most of the time the GTX 1070 is up by a minor amount of FPS (frames per second), but there are a few times where the GTX 980 Ti actually matches or surpasses the GTX 1070. And if looking at different resolutions, you will notice that by the time they are tested at UHD, the amount of FPS difference between the GTX 1070 and GTX 980 Ti is either gone or all but gone! Why should I pay (at least) another $100 for almost identical performance?

The only way the GTX 1070 would be more beneficial to me would be if I constantly used applications that could take advantage of the CUDA cores. While I do use at least one application where this could be beneficial, I do not presently use it enough to warrant an additional $100+.

Something else to add, that I did not know until I had my GTX 980 Ti, is that the AGA does not seem to allow for large overclocks like a desktop computer could. This means that paying more for a GTX 1070 would have been an even bigger loss if I had thought that an overclock might do far better than an overclocked GTX 980 Ti...

Bi Bi
So, even now, the GTX 980 Ti, despite NVIDIA slashing their prices, is still being sold for more than $350. The only other deal I saw that was similar to mine was for a used GTX 980 Ti on the same forum from where I bought my graphics card. This was probably only a couple weeks before I bought mine, and I do not believe it was even an EVGA card; meaning that the new owner cannot make any warranty claims if they should ever need to get the graphics card repaired.

I believe I got the best deal possible, and everything I have seen so far suggests that. Even when dwindling GTX 980 Ti supply gets to the price I paid (and remember mine was a total, market prices would not include tax), the amount of time that will have passed since obtaining my graphics card will be more than enough to be a major factor of what a great deal I got.