TweetFollow Us on Twitter

December 96 - Chiropractic for Your Misaligned Data

Chiropractic for Your Misaligned Data

Kevin Looney and Craig Anderson

Misalignment occurs when a program accesses data in a way that's not in sync with the processor's internal paths. This can slow down performance a little or a lot, depending on the CPU architecture. But finding these areas in code can be very difficult. We'll demonstrate the cause and cost of alignment problems and then show you a couple of tools you can use to detect them in your programs.

Sometimes Macintosh application performance is limited by architectural factors that can't be remedied, like the raw speed of the I/O or memory subsystem. But the programmer does have control over some factors that affect the speed of the memory subsystem and thus application performance -- such as how data is aligned and accessed within memory. By default, most compilers will do the appropriate alignment for PowerPC code. However, alignment options offered for backward compatibility with the 680x0 architecture can cause significant overhead.

Misalignment is a difficult performance problem to detect. Traditional debugging and performance tools typically don't help you find misaligned accesses. On top of this, misalignment problems manifest themselves differently on different CPU architectures.

In this article, we'll define misalignment, describe how it's caused, discuss the overhead penalties for accessing misaligned data on various microprocessors, and introduce some tools designed to aid in the detection of misaligned accesses in code. These tools accompany the article on this issue's CD and develop's Web site. Armed with these tools and what you learn in this article, you can perform chiropractic adjustments on your programs to solve their data alignment problems.


WHAT IS MISALIGNMENT AND WHY SHOULD I CARE?

A piece of data is properly aligned when it resides at a memory address that a processor can access efficiently. If it doesn't reside at such an address, it's said to be misaligned. In the PowerPC architecture, 32-bit and 64-bit floating-point numbers are misaligned when they reside at addresses not divisible by 4. Misalignment exceptions are taken based on the specific microprocessor.

Whether a data item is aligned depends not only on its address and the processor that's performing the access, but also on the size of the item. In general, data of size s is aligned if the least significant n bits of its address are 0, where n = log2(s). Hence, 1-byte items are always aligned, while 2-byte items are aligned on even addresses and 4-byte items are aligned if the address is evenly divisible by 4. This alignment policy is often called natural alignment and is the recommended data alignment for code to run well on all current and future PowerPC processors.

Accessing misaligned data can be quite costly, depending on the microprocessor your program is running on. We'll demonstrate just how costly in a minute, but in general, misaligned memory accesses take from 2 to 80 times longer than aligned accesses on 603, 604, and future PowerPC microprocessors. A misaligned access can require more time to perform for two reasons:

  • It may require two requests to the memory system instead of just one.

  • It may cause the processor to take an unaligned access exception, a costly penalty.


WHAT CAUSES MISALIGNMENT?

Misaligned accesses can involve variables located on the stack or on the heap. The type of compiler and the compiler settings that you use will determine whether misaligned accesses occur. Improper structure placement and incorrect pointer arithmetic can also cause misaligned accesses. These problems can be found with the tools and techniques described below, but this article generally focuses on alignment problems that aren't caused by programmatic errors.

Most compilers for the Macintosh allow you to choose among various alignment options. Some compilers default to 2-byte alignment so that data alignment in PowerPC code mimics alignment on the 680x0 processor. While using this option means that structures written to disk in binary format can be accessed easily by both architectures, it also permits alignment problems in the PowerPC architecture. Both improper structure padding and misaligned stack parameters can result in misaligned accesses.

IMPROPER STRUCTURE PADDING

When a float field occurs in a structure, improper padding by the compiler will cause the float field to be misaligned. The example in Listing 1 uses MPW's alignment pragmas to illustrate this.


Listing 1. An example of a poorly aligned structure

#pragma options align=mac68k 
typedef struct sPoorlyAlignedStruct {
   char      fCharField;
   float     fFloatField;
   char      fSecondCharField;
} sPoorlyAlignedStruct;

sPoorlyAlignedStruct      gPoorlyAlignedStruct;
#pragma options align=reset

In this example, a compiler that did no padding would align fFloatField on an offset of one byte from the structure's base address. Since compilers (and memory allocators) usually align the base of a structure on a boundary of at least four bytes (and multiples of four bytes), every access to fFloatField will cause a misalignment error. Also, fFloatField will be misaligned in statically or dynamically allocated arrays, since the lengths of structures are padded so that each structure that's an array element starts on a 4-byte aligned address.

A compiler with a 2-byte padding setting would align fFloatField on an even address, but this would still cause misalignment when that address isn't divisible by 4. Compilers using the mac68k pragma (as shown in Listing 1) cause 2-byte alignment, putting fFloatField on an even but often misaligned address for PowerPC processors.

A compiler with a 4-byte padding setting would always align the field properly.

MISALIGNED STACK PARAMETERS

Besides affecting the alignment of data in a structure, compiler settings can affect the way data structures are placed on the stack. Consider this function declaration:
void FunctionFoo (sPoorlyAlignedStruct firstParam, float floatParam)
In this example, the parameters are placed on the stack (even though PowerPC compilers use registers if possible). A compiler using a 680x0 2-byte padding option may align firstParam.fFloatField on an even address, but if the address isn't divisible by 4 this will cause a misalignment every time that parameter field is accessed within FunctionFoo. It won't, however, change the alignment of other parameters on the stack.

On the PowerPC processor, nonstructure parameters are usually placed in registers. There are no alignment problems when accessing registers.


THE COST OF MISALIGNMENTS

To demonstrate the cost of misalignments, we've written the code in Listing 2, which generates both aligned and misaligned accesses in the course of a million iterations. It accesses a byte array in different ways -- data writes of integers, floats, and doubles -- and at different offsets. In a portion of the code not shown, accesses are confined to within a single page of memory, and interrupts are turned off. Running this code enabled us to calculate the difference in performance between aligned and misaligned accesses. This code (with slight modifications for the various compilers) was compiled with the Symantec, MrC, and Metrowerks compilers. All compilers behaved similarly.


Listing 2. Generating accesses for comparison of access time

#define kNumAccessesPerCycle  200
#define kNumCycles            5000
// Number of total accesses = kNumAccessesPerCycle * kNumCycles
// 1000000                  = 200                  * 5000
#define kTableSize 1608 // Table size needed for 200 separate aligned
                       // accesses on the largest data type (doubles)

typedef enum ECType { eLong, eFloat, eDouble };

void main(void)
{
   double AlignedTimeFloat = AlignLoop(0, eFloat);
   double Misaligned1TimeFloat = AlignLoop(1, eFloat);
   double Overhead1Float =
       (((Misaligned1TimeFloat - AlignedTimeFloat) * 100)
           / AlignedTimeFloat);
   double avgOverhead1Float =
       (Misaligned1TimeFloat - AlignedTimeFloat)
           / kNumTotalAccesses;
   ...
}

// The function AlignLoop measures the time of a loop of "writes" to
// a byte array. The writes are either aligned or misaligned, based
// on the offset parameter, which should be between 0 and 7. The
// type should be eLong, eFloat, or eDouble.
double AlignLoop(short offset, ECType type)
{
   UnsignedWide   startTime, stopTime;
   double         start, stop;
   char           bytetable[kTableSize];
   long           j, k;

   // Get starting timestamp.
   Microseconds(&startTime);

   switch (type) {
      case eLong:
         {
            long *longPtr = (long *) &bytetable[offset];
            for (j = 0; j < kNumCycles; j++)
               for (k = 0; k < kNumAccessesPerCycle; k++)
                  longPtr[k] = 1;
         }
         break;
      case eFloat:
         {
            float *floatPtr = (float *) &bytetable[offset];
            for (j = 0; j < kNumCycles; j++)
               for (k = 0; k < kNumAccessesPerCycle; k++)
                  floatPtr[k] = 1.0;
         }
         break;
      case eDouble:
         {
            double *doublePtr = (double *) &bytetable[offset];
            for (j = 0; j < kNumCycles; j++)
               for (k = 0; k < kNumAccessesPerCycle; k++)
                  doublePtr[k] = 1.0;
         }
         break;
   }

   // Get ending timestamp.
   Microseconds(&stopTime);

   // Move the values to doubles.
   start = (((double) ULONG_MAX + 1) * startTime.hi) + startTime.lo;
   stop  = (((double) ULONG_MAX + 1) * stopTime.hi) + stopTime.lo;
   
   return stop - start;   
}

Table 1 shows the results. Overhead is calculated as the percentage difference between the time required for aligned and misaligned accesses. Our experiments showed that misaligned accesses at different offsets seemed to pay the same penalty (excluding cases where the two accesses required to retrieve the data straddle a memory page boundary, which is every 4K of memory).

Table 1. Misalignment overhead for basic data types, native PowerPC code. Access times are in (usec).

CPU and data Aligned total Misaligned total Overhead
access time access time
PowerPC 601 integers 113439 119573 5.4%
PowerPC 601 floats 63234 94505 50.0%
PowerPC 601 doubles 63251 113306 79.1%
PowerPC 604 integers 687 695 1.1%
PowerPC 604 floats 261 23753 9009.0%
PowerPC 604 doubles 262 22546 8509.5%

Note: Tests were run at 80 MHz on the PowerPC 601 and 132 MHz on the PowerPC 604.

Misaligned integer accesses result in a relatively small penalty for PowerPC 601, 603, and 604 CPUs. However, there's no guarantee that future microprocessors will provide hardware support for integer misalignment. Floating-point misalignments are severely penalized by the 604 implementation. In fact, while a misaligned access takes 1.5 times as long as an aligned access on the 601, it takes more than 80 times as long as an aligned access on the 604. The 601 pays a penalty only when an access crosses a page boundary (this will be verified later). Misaligned accesses to doubles on a 601 result in nearly double the overhead of misaligned accesses to floats. On the 604, however, doubles and floats suffer nearly the same overhead penalty. Misaligned accesses for doubles occur on any address not divisible by 8 on the 601, and any address not divisible by 4 on the 604. It's important to note that all memory accesses (aligned and misaligned) result in some timing penalty.

When we ran these experiments as emulated code (compiled for 680x0), float and double accesses showed no significant overhead (less than 8%). The 68040LC emulator doesn't do PowerPC floating-point loads/stores when processing floating-point data; it avoids alignment exceptions by doing integer emulation of a floating-point unit and loading and storing data 16 bits at a time.

Our code paints a worst-case scenario; worst case or not, the results indicate that there's plenty of motivation to avoid misaligned accesses in native PowerPC code. Perhaps the biggest problem facing the programmer, however, is the detection of these problems in application code. We'll look now at two tools that are useful for detecting and pinpointing alignment problems.


TOOLS FOR DETECTING MISALIGNMENTS

Apple's Performance Evaluation Group has developed two tools for detecting misalignments that cause exceptions:
  • PPCInfoSampler, a tool to detect high levels of misalignment exceptions over general application workloads

  • the Misalignment Instrument Library (MIL), a set of functions useful for profiling misalignments in specific regions of code
PPCInfoSampler is useful for determining whether your code has misalignment problems. If misaligned accesses are detected, the MIL can be used to pinpoint which parts of your code are causing the misalignments. We'll describe each tool in greater detail before discussing how to correct misalignments that you identify.

PPCINFOSAMPLER

PPCInfoSampler is a control panel that when activated records information about the PowerPC exception services and emulator at 100-millisecond intervals. The information recorded includes counts of mode switches, interrupts, misalignment exceptions, and page faults. See Table 2 for a list of PPCInfoSampler output categories. The output saved from PPCInfoSampler is in tab-delimited format and is best viewed from a spreadsheet program.

Table 2. PPCInfoSampler output categories

Output category Explanation
Time Delta (millis) Elapsed milliseconds since last sample of "exception services" registers
Microseconds time Microseconds (calculated from timebase) since PPCInfoSampler was enabled
Timebase Ticks A reading of the 64-bit timebase register
MixedMode switches Number of mode switches into PowerPC code
Data Page Faults Number of page faults
ExternalIntCount Number of external processor interrupts
MisalignmentCount Number of misaligned accesses that caused an exception
FPUReloadCount Number of reloads of the FPU register state
DecrementerIntCount Number of interrupts caused by the PowerPC decrementer register
EmulatedUnimpInstCount Number of instructions that are emulated in exception services
Timebase Ticks 68k Number of timebase ticks spent in 680x0 code
Timebase Ticks PPC Number of timebase ticks spent in PowerPC code
Level n Int Ticks Number of timebase ticks that expired per interrupt level
Level n interrupts Number of interrupts that occurred at each interrupt level

Note: With the exception of microseconds time, each measurement is per sample and isn't cumulative with the next interval.

To use PPCInfoSampler, you must first drop it into your Control Panels folder and reboot. The tool installs code in the system heap that waits for an action to occur. There are two ways to activate the sampling mechanism:

  • You can use the keyboard shortcut Command-Option-Z to start the sampler instantly at any time. When you start the sampler, an 8-pixel line will flash in the upper left corner of your main screen. It will continue to flash for each sample that's recorded. To stop the sampler, use the keyboard shortcut again.

  • You can use the control panel interface, as shown in Figure 1, to start, stop, and save a sample.

Figure 1. PPCInfoSampler control panel interface

For our purposes, let's focus on the number of misalignment exceptions. To determine in general whether you have misalignment problems in your application, think of the operations or "workloads" (such as saving to disk) that force your application to access many data structures. Run PPCInfoSampler during these workloads to determine whether any misalignments are occurring. Any misalignment count greater than 0 should be investigated and, if possible, corrected.

Table 3 shows an example of a misalignment count generated by PPCInfoSampler. The program that was being executed during this count displayed bursty misalignment characteristics. That is, during some 100-millisecond intervals no misalignments were happening; during other intervals, large numbers were happening.

Table 3. Sample output from PPCInfoSampler

Milliseconds Misalignment count
100 0
100 0
100 11652
100 43694
100 42931
100 43695
100 43679
100 43705
100 42942
100 31213
100 0
100 0
100 0
100 14510
100 44135
100 44667
100 44470
100 44347
100 44323
100 44303
100 6416
100 0
100 0

THE MISALIGNMENT INSTRUMENT LIBRARY

Where PPCInfoSampler allows you to detect misalignment activity in broad 100-millisecond intervals, the Misalignment Instrument Library (MIL) allows you to then make educated guesses about where to further instrument your code to pinpoint where the misalignment is happening. The MIL consists of two routines:
void initMisalignRegs(void);        // Initializes our misalignment
                                     // counter (do this only once)
unsigned long getMisalignments(void); // Returns the total number of
                                     // misalignment exceptions since
                                     // the initMisalignRegs call
With these calls to the MIL, you can profile strategic portions of your source for the application's workloads. Iteratively narrow the focus of your profile by moving the instrumentation in the code until you can determine where the misaligned accesses are and the structures that they're associated with.

Listing 3 is a sample of application code instrumented with the MIL calls. In this sample, we use the MIL to display the different floating-point exception-handling properties of the 601 and 604 CPUs.


Listing 3. Sample application code using the MIL

#define kNumAccessesPerCycle    200
#define kNumCycles  5000
#define kTableSize  804 // Table size needed for 200 separate aligned
                        // accesses on the largest data type (floats)
unsigned long gNumberOfMisalignments = 0; // Misalignments forced by
                                          // program
unsigned long gReportedMisalignments = 0; // Misalignments reported
                                          // by exception services
void main(void)
{
   float MisalignedTime = misalignLoop(false);
   printf(">*> Forced number of misalignments: %d\n",
            gNumberOfMisalignments);
   printf(">*> Reported number of misalignments: %d\n",
            gReportedMisalignments);
}

// The function misalignLoop measures the time of a loop of "writes"
// to a byte array. The writes are either aligned or misaligned,
// based on the align parameter.
double misalignLoop(boolean align)
{
   UnsignedWide   startTime, stopTime;
   float            start, stop;
   short            alignIndex = (align)?0:1;
   
   // · MIL instrumentation code: initialize misalignment
   // counter.
   initMisalignRegs();   
   // Get starting timestamp.
   Microseconds(&startTime);

   for (long j = 0; j < kNumCycles; j++) {
      char bytetable[kTableSize];
      float *floatPtr = (float *) &bytetable[alignIndex];
      for (long k = 0; k < kNumAccessesPerCycle; k++) {
         gNumberOfMisalignments++;
         floatPtr[k] = 1;
      }
   }

   // Get ending timestamp.
   Microseconds(&stopTime);
   
   // · MIL instrumentation code: get number of misalignments.
   gReportedMisalignments = getMisalignments();
   
   // Move the values to doubles.
   start = (((double) ULONG_MAX + 1) * startTime.hi) + startTime.lo;
   stop = (((double) ULONG_MAX + 1) * stopTime.hi) + stopTime.lo;
   
   return stop - start;
}

The results of running this code are shown in Table 4. There's a large discrepancy between the number of misalignments generated and the number reported by the MIL on the 601. The 601 architecture internally fixes float and double misalignments in hardware. However, the 601 can't fix misalignments across page boundaries, so it takes a misalignment exception. Thus, only those page boundary cases are reported. The 603/604 architecture doesn't handle misalignments in hardware, and it takes an exception in all cases.

Table 4. Number of misalignments generated and reported

CPU Number of misalignments Reported number of misalignments
PowerPC 601 1000000 120
PowerPC 604 1000000 1000000


NOW FOR THE CHIROPRACTIC ADJUSTMENTS

You've determined with the help of PPCInfoSampler that you have misalignment problems in your application. You've used the MIL to determine where the misaligned accesses are and the structures that they're associated with. Now it's time to think about these structures.

Is there a reason why they can't be naturally aligned (as we described early in the article)? If there are structures in the parameter block passed to a Toolbox call, fields within these structures may not naturally align, but this is something the programmer can't do much about until the system provides an alternate API. Perhaps there are binary data files created and accessed from 680x0 legacy code. Is it really necessary to still be supporting data files formatted to 2-byte alignment? Can you provide a version mechanism for your data files, such that newer versions write and read to natural PowerPC alignment? Ask these questions, and naturally align structures as much as possible.

You can ensure proper structure alignment by ordering the fields in your structures by hand, from largest to smallest, instead of relying on a compiler to pad the fields. This will require you to do more work but will remove reliance on any particular padding strategy.

Another possible scenario is the case where data files are shared across multiple platforms. Alignment strategies on Intel and other x86 processors aren't the same as on PowerPC processors. There are two possible approaches to this scenario, given an application on Windows and one on the Mac OS that share the same data files:

  • Try compiling structures on the Windows application with 4-byte alignment, according to PowerPC natural alignment. This is a "least common denominator" approach.

  • If rebuilding a Windows application isn't a viable option, load the data from disk and convert the data structures to an aligned structure that's used internally. The performance tradeoff depends on how often the misaligned structure is used.

IT'S THAT SIMPLE

Misaligned memory accesses can take a real toll on your application's performance, requiring from 2 to 80 times longer than aligned accesses on newer PowerPC CPUs. If you do what we've described in this article, you can detect and pinpoint misalignments and fix them so that your code will run efficiently now and on future processors (which won't include hardware to fix misaligned accesses for any misaligned data type) and won't be penalized by the lack of hardware support in future implementations of PowerPC architecture. Isn't it worth a few simple adjustments now to know that your code's future is secure?

KEVIN LOONEY (looney@apple.com) is a research scientist for Apple's Performance Evaluation Group, which does performance analysis of applications, systems, and hardware. He previously wrote performance and debugging tools as part of the Core Tools Group at Apple. Outside the confines of Apple, Kevin can be found moonlighting as a pianist/synthesist and Web designer. He mainly ponders two questions in life: why are things taking so long, and what would cause someone with a degree in artificial intelligence to study performance issues?*

CRAIG ANDERSON (c.s.anderson@ieee.org) was formerly a senior performance analyst for Apple's Performance Evaluation Group. He's now at a startup company. Before working at Apple, he spent many years researching and writing his best-selling work Improving the Performance of Bus-Based Multiprocessors. Craig enjoys cooking with Mollie Katzen and praticing katas with sensei Huber. He also takes delight in reading fine literature, such as The Gulag Archipelago, The Shipping News, and It's Obvious You Won't Survive by Your Wits Alone.*

Thanks to our technical reviewers Justin Bishop, Dave Evans, and Jim Gochee. The authors would like to acknowledge the help given by members of the Performance Evaluation Group (a subgroup of the Architecture and Technology Group in Apple's Hardware Division), including Marianne Hsiung, Tom Adams, and Scott McMahon. We'd also like to thank Jim Gochee for his toolsets and valuable insights.*

 

Community Search:
MacTech Search:

Software Updates via MacUpdate

GarageSale 7.0.8 - Create outstanding eB...
GarageSale is a slick, full-featured client application for the eBay online auction system. Create and manage your auctions with ease. With GarageSale, you can create, edit, track, and manage... Read more
Backblaze 5.0.0.116 - Online backup serv...
Backblaze is an online backup service designed from the ground-up for the Mac. With unlimited storage available for $5 per month, as well as a free 15-day trial, peace of mind is within reach with... Read more
Parallels Desktop 13.0.0 - Run Windows a...
Parallels allows you to run Windows and Mac applications side by side. Choose your view to make Windows invisible while still using its applications, or keep the familiar Windows background and... Read more
Mellel 4.0.0 - The word processor for sc...
Mellel is the leading word processor for OS X and has been widely considered the industry standard for long form documents since its inception. Mellel focuses on writers and scholars for technical... Read more
Adobe Muse CC 2017 2017.1.0 - Design and...
Muse CC 2017 is available as part of Adobe Creative Cloud for as little as $14.99/month (or $9.99/month if you're a previous Muse customer). Adobe Muse 2017 enables designers to create websites as... Read more
WhatsApp 0.2.5862 - Desktop client for W...
WhatsApp is the desktop client for WhatsApp Messenger, a cross-platform mobile messaging app which allows you to exchange messages without having to pay for SMS. WhatsApp Messenger is available for... Read more
WhatsApp 0.2.5862 - Desktop client for W...
WhatsApp is the desktop client for WhatsApp Messenger, a cross-platform mobile messaging app which allows you to exchange messages without having to pay for SMS. WhatsApp Messenger is available for... Read more
Things 3.1.3 - Elegant personal task man...
Things is a task management solution that helps to organize your tasks in an elegant and intuitive way. Things combines powerful features with simplicity through the use of tags and its intelligent... Read more
BetterTouchTool 2.292 - Customize Multi-...
BetterTouchTool adds many new, fully customizable gestures to the Magic Mouse, Multi-Touch MacBook trackpad, and Magic Trackpad. These gestures are customizable: Magic Mouse: Pinch in / out (zoom... Read more
Things 3.1.3 - Elegant personal task man...
Things is a task management solution that helps to organize your tasks in an elegant and intuitive way. Things combines powerful features with simplicity through the use of tags and its intelligent... Read more

KORG iMono/Poly (Music)
KORG iMono/Poly 1.0.0 Device: iOS Universal Category: Music Price: $19.99, Version: 1.0.0 (iTunes) Description: *** Special Sale for a limited time to celebrate the debut of KORG iMono/Poly (33% OFF) until Sep 30! *** Reviving a... | Read more »
Super Phantom Cat 2 beginner's guid...
Super Phantom Cat 2 presents a whole new world of fun platforming challenges and perplexing puzzles. It's a well-designed platformer with a bright, neon aesthetic that brings the genre up to date. [Read more] | Read more »
Shadow Fight 2 Special Edition (Games)
Shadow Fight 2 Special Edition 1.0.0 Device: iOS Universal Category: Games Price: $4.99, Version: 1.0.0 (iTunes) Description: ** New story chapter! **** No Ads! **** No energy! ** The best fighting series on mobile has returned and... | Read more »
4 RPGs like Final Fantasy XV that deserv...
Square Enix announced another Final Fantasy XV spin-off today - Final Fantasy XV Pocket Edition. This mobile, episodic version of the hit RPG gives the game a chibi-fied makeover. The first episode will be free, followed by 9 more premium episodes... | Read more »
Guild sieges and soul gems in latest upd...
Webzen’s MU Origin hit app stores last year, giving fans of fantasy hack-n-slash MMOs like Diablo a new fix to fixate on. This latest update introduces a competitive guild battle, a fresh dungeon challenge, a mini-game and some elemental gems to... | Read more »
Little Red Lie (Games)
Little Red Lie 1.0 Device: iOS Universal Category: Games Price: $4.99, Version: 1.0 (iTunes) Description: ARE YOU MORE AFRAID OF POVERTY THAN DEATH? Little Red Lie is a narrative-focused, interactive fiction experience that reduces... | Read more »
You can now apply to be Clash of Clans...
Earlier this month, word got out that the Builder, the trusty handiman who tirelessly built every single building inevery singleClash of Clansbase had called it quits. Sick of seeing his work destroyed endless, the Builder has set out for our world... | Read more »
Meshi Quest beginner's guide - how...
Meshi Quest is Square Enix's newest free-to-play release, and it's a real charmer. You start off as the head of a sushi restaurant, upgrading your food and equipment as you serve visitors heaping helpings of your delicious meals. As you progress,... | Read more »
BUST-A-MOVE JOURNEY (Games)
BUST-A-MOVE JOURNEY 1.0.0 Device: iOS Universal Category: Games Price: $4.99, Version: 1.0.0 (iTunes) Description: BUST-A-MOVE Features:- Shoot bubbles and match 3 or more bubbles of the same color to make them pop!- Complete your... | Read more »
The best card combos in Clash Royale
Clash Royale is all about building a deck of units that synergise well. To help you get off to a flying start, we've put together a list of unit combinations that are incredibly effective. Looking for some choice 2v2 combos? Check out our guide. [... | Read more »

Price Scanner via MacPrices.net

Clearance 2016 13-inch MacBook Airs, Apple re...
Apple has Certified Refurbished 2016 13″ MacBook Airs available starting at $809. An Apple one-year warranty is included with each MacBook, and shipping is free: – 13″ 1.6GHz/8GB/128GB MacBook Air: $... Read more
2017 13-inch MacBook Airs on sale for $100 of...
B&H Photo new 2017 13″ MacBook Airs on sale today for $100 off MSRP, starting at $899: – 13″ 1.8GHz/128GB MacBook Air (MQD32LL/A): $899, $100 off MSRP – 13″ 1.8GHz/256GB MacBook Air (MQD42LL/A... Read more
Sale! 13-inch 2.3GHz MacBook Pros for $100 of...
B&H Photo has 13″ 2.3GHz MacBook Pros in stock today and on sale for $100 off MSRP including free shipping plus NY & NJ sales tax only: – 13-inch 2.3GHz/128GB Space Gray MacBook Pro (MPXQ2LL... Read more
2016 MacBook Pros, Apple refurbished, availab...
Apple has Certified Refurbished 2016 15″ and 13″ MacBook Pros available starting at $1189. An Apple one-year warranty is included with each model, and shipping is free: – 15″ 2.7GHz Touch Bar Space... Read more
Apple offers Certified Refurbished iPhone 6s...
Apple has Certified Refurbished unlocked iPhone 6s’s and 6s Plus’s available starting at $449. An Apple one-year warranty is included with each phone, and shipping is free: – 16GB iPhone 6s: $449, $... Read more
Apple offers Certified Refurbished Pencils fo...
Apple has Certified Refurbished Apple Pencils available for $85 including free shipping. Their price is $14 off MSRP, and it’s the lowest price available for a Pencil. Read more
2016 15-inch 2.6GHz Touch Bar MacBook Pro ava...
B&H Photo has clearance 2016 15″ 2.6GHz MacBook Pros in stock today and on sale for $500 off original MSRP. Shipping is free, and B&H charges NY & NJ sales tax only: – 15″ 2.6GHz Touch... Read more
21-inch 2.3GHz iMac on sale for $999, save $1...
Amazon has the new 2017 21″ 2.3GHz iMac (MMQA2LL/A) in stock and on sale for $999.99 including free shipping. Their price is $100 off MSRP, and it’s the lowest price available for this model. Read more
Free Instant Translator 2.0 App For iOS Relea...
Mobile application development company, Neoappz has announced the release and immediate availability of Instant Translator 2.0 for iOS devices. Instant Translator is a user-friendly application which... Read more
2017 15-inch MacBook Pros on sale for $200 of...
Amazon has 2017 15″ MacBook Pros on sale for $200 off MSRP. Shipping is free: – 15″ 2.8GHz MacBook Pro Space Gray: $2199.99, $200 off MSRP – 15″ 2.8GHz MacBook Pro Silver: $2296, $103 off MSRP – 15″... Read more

Jobs Board

Development Operations and Site Reliability E...
Development Operations and Site Reliability Engineer, Apple Payment Gateway Job Number: 57572631 Santa Clara Valley, California, United States Posted: Jul. 27, 2017 Read more
Frameworks Engineering Manager, *Apple* Wat...
Frameworks Engineering Manager, Apple Watch Job Number: 41632321 Santa Clara Valley, California, United States Posted: Jun. 15, 2017 Weekly Hours: 40.00 Job Summary Read more
Development Operations and Site Reliability E...
Development Operations and Site Reliability Engineer, Apple Payment Gateway Job Number: 57572631 Santa Clara Valley, California, United States Posted: Jul. 27, 2017 Read more
Frameworks Engineering Manager, *Apple* Wat...
Frameworks Engineering Manager, Apple Watch Job Number: 41632321 Santa Clara Valley, California, United States Posted: Jun. 15, 2017 Weekly Hours: 40.00 Job Summary Read more
*Apple* Retail - Multiple Positions - Apple,...
Job Description: Sales Specialist - Retail Customer Service and Sales Transform Apple Store visitors into loyal Apple customers. When customers enter the store, Read more
All contents are Copyright 1984-2011 by Xplain Corporation. All rights reserved. Theme designed by Icreon.