Eight computers are better than one

Parallel CPU

Use All of your Computer - Not just 1/8th

Multi Core + Multi CPU

Decades ago in the 1990's all desktop computers had only one processor in their CPU chip. Today's computers have many processors in their CPU chip. Some desktops even have multiple CPUs with many processors in each. That gives your modern computer the power of many old-fashioned, single-processor computers, but only if you run multicore parallel software like Radian.

Non-parallel software still runs all computers as if they were old-fashioned 1990's computers with only one core per task. No matter how modern your computer, other software imposes 1990's limits on you. Your computer could have 20 processors in the CPU chip but non-Radian software will use only one of those cores per task - just as if your computer was a 1990's computer. The chart shows how Radian gains proportionately more performance with more cores. The more cores you have, the bigger the gain. With eight cores Radian is way more powerful than single core software. With sixteen cores nothing else comes close.

The graphs below show actual Radian performance in Windows 10 Resource Monitor using an inexpensive, eight hypercore, Intel Core i7 processor, with and without parallel processing of an image transform. Radian always uses all cores for a big process, but other software can't do that. It might hop around from core to core, but just like the 1990's it never uses more than one core.

Radian: Full Power

Radian is Parallel

Others: 1/8th Power

Old-style software is not Parallel

What do you call software that only uses 1/8th or 1/20th of your computer? If you've just spent a few thousand dollars for your computer you might call it some angry names. We call it non-Radian software because only Radian uses all of your computer all of the time to do your work faster and easier. Non-Radian software that uses just one core out of eight is wasting over 87% of the computing power of your CPU.

Compare CPU Performance

Compare Radian's effectiveness to other software running right out of the box, with no special software engineering efforts required, just routine, point-and-click, always on, fully automated use on the desktop by normal people with average DBMS and GIS skills:

Inexpensive Desktop - 8 Cores (AMD FX-8300, Intel Core i5-7500)

Package GPGPU CPU Cores Cores Used % Used % Wasted
Not Parallel No 8 1 12.5% 87.5%
Radian Yes 8 8 100% 0%

Multicore hardware is now very affordable: A $100 AMD FX-8300 provides eight CPU cores while a $250 Intel Core i5-7500 provides eight CPU hypercores. Even better, the $100 AMD chip with eight cores running Radian will crush a $1600 chip running non-parallel software. You get all that with Radian even without using a GPU.

Professional Desktop - 20 Cores (Intel Xeon, Intel Core i7-6950X)

Package GPGPU CPU Cores Cores Used % Used % Wasted
Not Parallel No 20 1 5% 95%
Radian Yes 20 20 100% 0%

Professional desktop machines using cutting edge multicore processors cost more but are still affordable: a 16 hypercore AMD Ryzen is only $320. Intel's Core i7 chips provide up to 20 hypercores. A pair of surplus Xeon CPUs at only $50 per chip provides a total of 32 hypercores. Running Radian on 32 inexpensive hypercores will easily outperform far more expensive systems running non-parallel software. Add a $50 GPU and you often can run hundreds of times faster than others.

Use All Of Your System's Power...

Does it make sense to spend a lot of money on a computer and then throw away most of what you bought?

  • One Cylinder out of Eight? - Would you buy an automobile with an eight cylinder engine and then fill it with fuel that allowed only one cylinder to work while seven cylinders did nothing?
  • One Hour a Day? - Would you hire somebody that worked only one hour out of an eight-hour day but got paid for all eight hours?
  • Why do that with software? If you wouldn't do that to your car, and if you wouldn't hire such a lazy worker, why do that to your computer? If your computer has eight CPU cores, why use software that allows only one core to run a process while seven cores stay idle?

... Not Just a Little

Software that can run only one core out of eight cores in your system wastes seven eighths of your system - almost all the power in your system. Have you wondered why non-Radian packages are so slow with demanding jobs? One reason is they fail to use most of the computing power of your system.

The tables above show how the ratio of effective use to wasted power is even worse with newer computers. The more modern your processor the more wasteful non-Radian software will be. If you are running a modern processor like an Intel Core i7-6950X you have twenty hypercores. Radian uses all 20 hypercores. Automatically. Non-parallel software will use only one hypercore and will waste 95% of the computer's power. Only Radian automatically uses all of the computer.

Even Faster: Multiple CPUs

If your system has multiple CPUs Radian will use all of the cores in every CPU. Professional users in GIS and spatial data engineering can deploy desktop systems with two CPU sockets. Using a 16 or 20 hypercore Intel Core i7 CPU in each of those two sockets will provide a total of 32 or 40 hypercores on the desktop. Radian will use them all for astonishing power.

Prices are plummeting on many-core CPUs. With 16 hypercore AMD Ryzen CPUs appearing in the $300 range a 32 core system using two such chips on the motherboard becomes affordable for most professional users. $600 for 32 cores is way cheaper than $2200 or $3200. If you run Radian you can enjoy blistering performance at absurdly low price.

Parallel Processing - The Real Thing

Superior spatial data engineering requires parallel power, and real parallelism cannot be glued on to non-parallel code. It has to be built in from the ground up.

Radian was created from the ground up as an always-on, fundamentally parallel system, with parallel data systems feeding massively parallel computation. The Radian database engine is a parallel engine and Radian SQL automatically runs parallel by default. Every bit of Radian was designed and coded for fully parallel work, including all infrastructure, hundreds of functions exposed for users and thousands of functions utilized internally. Zero weak links.

"Radian Studio did in two hours what Manifold 8 has yet to do in 85 hours!" - User post on beta forum, regarding a UNION of 100 tables involving all real estate parcels in a major US State, a very big job.

All of Radian is a true multicore, parallel processing system. Radian will automatically utilize multiple processors and multiple processor cores by parallelizing a task into multiple threads for execution on all of the cores in your system. Given hyperthreading plus multi-core CPUs, even inexpensive desktops will have 8, 16, 32 or even more CPU cores available. Radian will use them all, as the Resource Monitor graphs show above.

In addition to parallel processing using multiple CPU cores Radian also automatically launches massively parallel multiprocessing utilizing GPUs, potentially launching tasks on thousands of GPU cores at once for true supercomputer computational performance, far beyond what can be achieved with multicore CPU alone and often hundreds of times faster than other software. See the GPU Page for more.

Radian automatically parallelizes and dispatches hundreds of functions to GPGPU, with automatic fallback to parallelized tasks dispatched to multiple CPU cores if a GPU is not available. It's all automatic and totally transparent with zero user effort required to use. Just do what you do and Radian handles all the parallel wizardry.

No Parallel Coding Required

You don't need to learn any parallel programming. Radian does it all for you with automatic parallelization. Write the same SQL you already know and Radian automatically parallelizes it for execution. Use Radian functions in your scripts (Radian includes multithreaded V8 JavaScript plus use of 10 other languages) and those run parallel as well. Point and click with hundreds of Radian templates without any programming at all and those all run parallel too. Click open immense displays and Radian parallelizes the rendering, parallelizes re-projections on the fly and everything else. Radian runs parallel on however many CPU cores you have. Automatically. No parallel code required.

Cores and Hypercores

Cores and hypercores mean almost the same thing: both refer to using multiple physical or logical processors on the same silicon chip to give the power of multiple computers within one chip. Cores usually means extra silicon circuitry, multiple copies of a processing unit. Hypercores is an Intel word for hyperthreading that gives a similar effect by leveraging idle circuitry to provide the horsepower of extra processors. 20 hypercores are not as powerful as 20 physical cores, but they can be close enough to count them the same. Windows Resource Manager counts them the same when it reports the number of CPUs in your system. Some chip vendors prefer the terms cores and threads to avoid using Intel's proprietary word. No worries - Radian can use them all for true multicore parallelism.

When "Multi Core" Means "One Core"

Do not be confused by software that claims to be "multithreaded" or "multicore" but uses only one core for a task. Much software advertised as "multithreaded" is not parallel. It does not run a task in parallel on many cores like Radian. Instead, it only uses one core for that task.

A well-known GIS package, for example, that is not parallel nonetheless uses "multithreaded" to describe itself in sales pitches. What they mean by "multithreaded" is the package can run one geoprocessing task in background on a single core, while the interactive user interface in foreground does not freeze up. Any additional tasks must wait for the first one to finish. Only then the next task can run, but again only on a single core. In that package big tasks run on a single core only.

That is not multicore parallel processing like Radian. If you have sixteen cores that well-known company's "multithreaded" software will waste fifteen cores, painfully grinding through the geoprocessing task using only one core. Radian uses all sixteen of the cores together at the same time, a huge difference.

Some vendors who have failed to create a parallel system try a different sales tactic. They implement a handful of functions with parallel capability or they call another package, perhaps on a cloud somewhere, and claim that capability as their own. Some will say "Yes, we are multicore, too..." even if less than 1% of their package is parallel and the overall effect is near zero.

Such tactics still waste most of your computer. First, the package itself remains a non-parallel bottleneck and second, implementing a few functions out of hundreds means that almost always your computer will be wasted. As one such vendor explains, "The use of multiple cores only applies to certain operations/functionalities [...] which is why you may not have noticed much difference in your CPU usage."

Do not be fooled by "multithreaded" or "multicore" packages that run tasks on only one core. Only Radian always runs parallel, automatically chopping up large tasks to run on all cores in your computer. No other desktop GIS or spatial engineering tool can do that.


Suggestions

Suggestions to improve Manifold are always welcome. Please see the Suggestions page for tips on making effective suggestions.


Buy Now via the Online Store

Buy Manifold products on the Online Store. The store is open 24 hours / seven days a week / every day of the year. Orders are processed immediately with serial number email sent out in seconds. Use Manifold products today!

 Click to begin shopping


Radian

Terrain elevations in Charles County, Maryland, from an ESRI ArcGIS REST web server operated by Salisbury University on behalf of Maryland government. Colored on the fly with an elevation palette and hill shaded by Radian.

About Manifold

Manifold products deliver quality, performance and value in the world's most sophisticated, most modern and most powerful spatial engineering products. Total integration ensures unbeatably low cost of ownership.