Related InformationRelated Information

Contact UsContact Us

Keep up-do-date with all things Grapeshot, Click Here to register your details and we'll keep you informed.

Tel:
+44 (0)1223 311319
Email:
info@grapeshot.co.uk

View the Showreel

BenchmarksBenchmarks

Stop Watch
Grapeshot is fastest on the clock
  • "Grapeshot is outstanding"
  • "Grapeshot ...in our experience market leading"

So many search engines to choose from...

There are over 300 search vendors our there. Some do just Windows systems, other just Java. Some have standard keyword search, and others like Grapeshot have many more advanced capabilities. What sorts the industry leaders from the rest of the pack is scale - the ability to execute very fast searches on large volumes of documents.

Champion Search

Cups
Grapeshot's champion performance

"Let me start of my thanking you for the opportunity to evaluate Grapeshot. It has been fascinating and we feel that you have a uniquely powerful product.. Our conclusions are that the performance of the engine is outstanding and certainly in our experience market leading."

3rd party evaluator comparing Grapeshot with many search engine vendors' products and open source Lucene.

Size of Software Footprint (MB) and Corresponding Memory Overheads

Grapeshot packs in latest developments

Grapeshot is the newest search technology (built 2002-2005), but created by a team with strong search pedigree and deep experience in the field of search and Information Retrieval (1980-2006).

Recent benchmarks show that Grapeshot can technically outperform in terms of speed, advanced bayesian functionality and sheer compactness of code footprint.

Typical Grapeshot Benchmark

1 Intel/Linux
1 2.4GHz CPU
2 GB RAM
7200rpm Hard Drive
Documents Indexed Per Second 210
Query Response Time 0.004
Number of Documents 3 million
Concurrent Queries 60

Comparison

Of all search vendors, Grapeshot is probably most similar to search engine vendors which use Bayesian methods, in terms of how "concept" searching is performed. Grapeshot deploys recent algorithms to achieve speed and performance using cheap hardware.

Comparison Item Grapeshot
Queries Per Second Per CPU 200
Response Times (seconds) 0.004
Data Indexing Rate per hour per CPU 5GB

Source: Grapeshot benchmarks using Moreover news, courtesy of Verisign Inc., on hardware costing less than $1000.

Grapeshot is confident it can outstrip many search products in terms of speed performance over millions of documents, based on customer feedback. Put us to the test! Our cost of purchase and ownership is often 10X less too!

Need for Speed

At the end of the day application builders need one feature above all else - fast robust performance software. Your application needs to work but it also needs to really scale.

Martin Porter, Grapeshot's creator, has many search engines like Lucene, Google, Microsoft, IBM and corporations the world over using his Porter Stemmer algorithms. His algorithmic approach has produced speed and scale. In much of information retrieval today there is a focus on hand built thesauri and semantic webs which actually create some slow and cumbersome systems that cannot scale as the language and domain focus makes them too specific and requiring word look-up processes.

Grapeshot team also has experience of building large indexes (>500 million documents) of the internet back in 1999, on low-end Linux equipment, when Alta-Vista was pumping large Digital infrastructure to achieve just a 300 million index.

Delivery of Speed

Grapeshot can index over 200 XML documents per second on a cheap single 2.4 GHz processor. Grapeshot can service at least 65 transactions per second on a single Java JNI thread - in heavy load situations where there are 14 million searches conducted per day.

Distributed Search

Grapeshot's distributed search architecture can also merge 1000 or more indexes, each with a capacity for 2 to the power 47 documents, together into one unified search. This has tremendous impact on how indexing is performed, offering a distributed design with indexing happening more locally to where data is changing or updating.