Showing Results: Most Recent
This review is from: FREMO P130 13000mAh Power Bank External Battery Charger for iPhone, iPad Air, mini, Galaxy S5, Note, Galaxy Tab, Nexus, HTC One, One 2 (M8), PS Vita and other Smartphones and Tablets (made by SCUD)
Pros: Pretty small, does it's intended purpose, simplistic, and stays cool to the touch. The metal shielding (while looking cool) also serves as decent protection should you drop it.
Cons: Can't think of anything really.
Other Thoughts: Note this power bank isn't (easily) user-serviceable. Opening it up requires you to physically rip off either the front or back white covers.
This power bank at full-charge kept a Google Chromecast powered for a little over 33 hours on the 1A USB port.
Took around 8-9 hours to fully charge after being completely drained (charged using the OEM Nexus 7 2013 ASUS charger).
The power bank remained cool to the touch during all periods of use and charging.
Although the power bank does it's intended purpose well, I think it would have been pretty cool to incorporate a Qi wireless charger. Can easily sit a device on-top of the battery. Might also be nice to have some sort of diagnostic statistics available (things like battery health, temperature, charging cycles, etc), either via USB, or NFC. Just to be clear, the P130 in it's current state does not support Qi wireless charging nor give diagnostic information.
Also if you need access to more ports for some reason, an idea could be to hook up a USB hub.
This review is from: ASRock Fatal1ty H97 Killer LGA 1150 Intel H97 HDMI SATA 6Gb/s Intel Motherboard
Pros: Works well, BIOS is reliable, and has features not available to the stock H97 chipset.
This motherboard also has unlocked MSRs, which is nice (relevant to OS X).
Was able to overclock my i5-4670K without issue.
The VRM and Southbridge heatsinks use screws, which makes taking them off and modifying easy (no dealing with those annoying 1-time use push pins).
Works great with Windows 8.1 x64, Linux (Ubuntu 14.04 x86_64 with kernel 3.15.1), and OS X Mavericks (10.9.3; you will need to mess around with kexts specifically for PS/2 (can use VoodooPS2), the Ethernet, and Audio (can use VoodooHDA), possibly even PM if using SpeedStep). All operating systems work fine with UEFI, and CSM disabled. Windows 8.1 works fine with Ultra Fast Boot, and Ubuntu works fine with Fast Boot. Works fine with the Clover (OS X) and Grub2 bootloaders as well.
This motherboard even has dual-BIOS chips. I wouldn't assume you are totally 100% safe from bricking the motherboard with flashing a BIOS (although it's possible you just may well be unless you flash both BIOS chips with broken firmware), but you should be far safer.
Cons: There is no way to select a particular EFI file to boot from within the BIOS GUI interface. I was unable to boot Clover with just using a EFI/BOOT/BOOTX64.EFI file on the HDD, but was able to when it was on a flash drive. I had to use the bootmgfw.efi trick to get an entry showing initially (Windows Bootloader).
The EFI Shell also appears to be a bit incomplete (typing help shows a bunch of null commands).
The chipset is a bit strange, but I do applaud ASRock for making it better. Basically, there is the Z97 chipset, and the H97 chipset, which is basically a stripped-down version of the Z97. As to why Intel launched such a chipset, I don't know, but in any case, ASRock did re-implement plenty of the stripped features (most notably Crossfire and overclocking). Re-implementing such features though doesn't come without downsides though; specifically, there are strange expansion slot card speed conditions. Basically, a 2nd GPU in the PCI-E slot will out-of-the-box run slower than the 1st slot, and if another expansion card is inserted elsewhere, the PCI-E slot speed gets cut in-half (from x4 to x2). Using a single GPU is fine though, and will have no problems running at the advertised PCI-E 3.0@x16 specification.
Other Thoughts: As most people claim, the Killer NIC isn't as great as it sounds. The software in-particular on Windows isn't something I would vouch for personally. However, you should be aware that the E2200 chip is just a slightly-modified AR8161 Atheros chip, and thus, you can use it's driver. Under Windows, you have to manually force the installation of that driver, but based on testing and feedback from others, it works fine, and on-top of being far less bulky, it may even be more stable than the Killer-specific driver.
The unlocked MSRs note is relevant if you plan to install OS X with this motherboard. With a motherboard with locked MSRs, OS X (Mavericks specifically) will reboot early in the booting process when trying to access a specific CPU register. This motherboard (along with certain other Fatal1ty boards) has unlocked MSRs, which means no boot-time kernel patching is needed. It's unknown if Windows or Linux benefits from this though, but it's still a nice touch.
There is a Non-Z OC feature for this motherboard. I however didn't need to enable it in order to overclock my 4670K. According to an article however, it seems this feature is needed to overclock K and non-K CPUs on non-Z chipsets (like the H97 chipset this motherboard uses).
This motherboard also worked fine with the CM Hyper 212 Plus cooler I have.
The USB ports on this motherboard are a bit interesting (in a good way). First off, the labeled USB 3.0 ports work even though USB 3.0 is disabled in BIOS. However, the ports will function as 2.0 ports. Another thing is the USB Compatibility Path setting in BIOS. Under normal circumstances, the setting should be disabled, and based on my testing, works fine with just about everything, except for VMware Player (likely other VMware virtualization products and other similar software). The Player would not allow me to pass USB devices over to the Guest OS until I enabled the patch setting. Normal USB usage still worked fine though.
There is also a red light under the PCIE2 slot. I imagine this just lets you know the motherboard has power and is turned on, but I couldn't find any information about it.
From a cold boot, when pushing the power button, the system turns on briefly, and then turns back off, and then powers on for real. The reasoning for this is unknown in my situation, but perhaps it has to do with the CPU and overclocking.
When using Linux, kernel 3.13 (included in Ubuntu 14.04), the onboard audio was glitchy and basically unusable. Upgrading to 3.15 fixed the issue though. I'm unsure how kernels in-between would work with audio.
Although Crossfire may not be the most ideal experience with this motherboard, you could try another interesting feat. If using a compatible CPU (you need VT-d for this, and all the K CPUs from Intel aside from the upcoming Devil's Canyon CPUs lack this feature), you could attempt to use GPU-passthrough with virtual machines.
Pros: Performance is pretty good.
I was able to reach between 50-70 FPS in Guild Wars 2 while standing in Divinity's Reach. Compared to my old CPU (listed in other thoughts), that's a huge and impressive improvement (20-30).
I get 40+ FPS in most cases with Grand Theft Auto IV with settings on high, 100 vehicle density, 50 everything else. Had to drop to medium/low and to 1 on the draw distances to get comparable framerate on my old CPU.
Everything else I played (Path of Exile, Dota 2, WarCraft III, Dynasty Warriors 8) did see an improvement as well, but the performance of the old CPU wasn't too bad either to make it really noticable.
Windows 8.1's performance is noticeably better as well.
Cons: Runs a bit hot. Temperatures while running prime95 at 4GHz @ 1.2V climb upwards to around 80C. Idles around 40C. The temperatures spike up pretty quickly too (from 40C to high 60s the second I start the stress test).
Doesn't support VT-d. If you're familiar with Intel, this shouldn't be a surprise though (unlocked CPUs usually lack certain virtualization features).
Other Thoughts: Firstly, if you are considering this CPU at the time of writing, and have a Z97 or H87 chipset, I may suggest waiting about a week and getting the 4690K for around the same price. Should also note the 4690K has advertised support for VT-d too, and is rumored to be using a better thermal compound in-between the CPU die and IHS (for better temperatures).
I upgraded from a Phenom II X3 720 (unlocked to X4 @ 3.2GHz) and an ASRock 970 Extreme3 R1.0 board, so in that regard, the 4670K is a really nice and noticeable upgrade.
I overclocked my 4670K to 4.0GHz across all cores, with a cache ratio of 40. Vcore voltage is 1.2V. YMMV.
My environment consists of an ASRock Fatal1ty H97 Killer, a CM Hyper 212 Plus (using Arctic Silver 5 thermal compound with the dot in the center method), and a hec 585W PSU (about 3 years old). That PSU in-particular only has a 4-pin CPU connector, which I can confirm is enough to reach the above clock speeds (anything higher isn't ideal temperature and stability-wise in my setup). My GPU is a MSI R7850 PE 2GD5/OC @ 1GHz core clock, at 1600x900@75Hz.
So about the lack-of VT-d, it seems relatively common that Intel strips certain virtualization features from their unlocked CPUs. I was interested in trying out some GPU-passthrough and virtualization. Considering my old setup met the requirements to do the passthrough (IOMMU), I thought it wouldn't be a problem with way-newer Intel hardware. Needless to say, I was rather surprised by the lack-of that feature. As for why Intel does this, it seems to be mostly speculation, but an idea is to just keep their non-unlocked CPUs relevant.
And as for a temperature note, it seems Intel chose to use a cheap thermal compound in-between the CPU die and IHS (the top cap you sit the heatsink on; note the die is significantly smaller than that area). Previous CPUs by Intel used solder, which transfers heat from the die to the IHS far more efficiently. There's no official-word for the reasoning as far as I know, but there are two ideas:
- Due to the die manufacturing process, using solder and having temperature changes would cause the die itself to crack under stress from typical usage scenarios (would be bad for PR and the RMA department).
- It's cheaper to use thermal compound over specialized solder (this wouldn't surprise me coming from Intel).
And on-top of the cheap thermal compound, the IHS is secured to the PCB via thick glue. Because of this, the IHS and die don't come in-contact with each other as close as they can, which then leads to poor heat transfer.
You can attempt to fix the temperature situation yourself via "delidding", but be very aware of the risks of doing so (one wrong scratch on the CPU die and it's likely done), along with accepting loss of any warranty. It cannot be stressed enough; do your research before even thinking about trying delidding! The risk is well worth the reward in some cases though (significantly lower temperatures, thus more O