Benchmarks - VaR

In terms of the time it takes to parse, load, and aggregate VaR data, ActivePivot has reduced the time it takes to perform this process to the point that it easily fits within a bank's end-of-day window and now makes intra-day VaR a feasible task.

Using the ActivePivot performance/memory figures from a recent project Quartet FS did for a very large global bank for their global VaR reporting requirements; on an 8-core, Intel Nehalem server with 64 GB RAM, ActivePivot was able to load 8 million trades with an average of 4 risk types each, with 1,000 simulated P&Ls per trade, for ten days - or in others words, 320 billion P&Ls.

ActivePivot parsed, aggregated and loaded into its cube these 320 billion P&Ls in 80 minutes with the memory usage being only 42 GBs. Therefore, ActivePivot was able to load more than 66,000 complex objects (a trade with all its dimensions and measures (including the 1,000 P&Ls)) per second into the cube. Each subsequent day's new VaR data (32 billion P&Ls per day) was loaded within 8 minutes.

The cube used at this bank has 18 dimensions, 36 levels, with 10 measures - among them is the vector of 1,000 P&L simulations - and allows the users to "slice & dice" through these dimensions using Excel or our "ActivePivot Live" browser-based user interface, with sub-second response time - seeing VaR at whatever level is necessary.

Please take a look at our new online browser-based ActivePivot VaR demo. It only takes a minute to register for access to it. Click here to register for our online VaR demo.

Read below for the main points of the online demonstration:

• The "ActivePivot Live" user interface is browser-based which means zero deployment. It can be extended by your IT department so new features and the "look & feel" itself can be enhanced by your own organization. Have a look at the info written up for each section as signified by the little "info button".
• On the right side of the screen is the typical pivot table functionality to create your own view by first picking the cube to look at and then dragging and dropping the measures and the dimensions into the columns, rows, and/or filters below. Then you can click on the filters at the top in the middle and do multi-selects, etc. Also, you can click on the columns and rows in the middle to drill-through the data and then double click any aggregated value and see the trade level results in the frame below. Once you create a view you like, right click in the middle and save it as a bookmark on the left to use again.
• On the left side of the screen are the bookmarks, which can be shared and are collaborative (like Google Docs). These are saved views that users can just click on and the current numbers in the cube are displayed for those pre-saved selection criteria.
• Double click one of the values in the pivot table to see a detailed list of data in the Drill-through Blotter which contributed to that cell's value.
• Right click on a value in the pivot table for added functionality - such as a graphical VaR histogram (historical VaR, stress VaR, or independent scenario VaR).