Behind the scenes at Labs
Showing results for 
Search instead for 
Did you mean: 

INFOGRAPHICS: A peek under the hood of The Machine full-scale prototype


Blog_650x550.pngCome take a "peek under the hood of The Machine full-scale prototype" at HPE Discover, June 5-8. 


T12965 – A peek under the hood of The Machine full-scale prototype

Date: Wednesday, June 7th   

Time: 10:00-10:30am

Moderator: Martina Trucco

Panelists: Andrew Wheeler, Lisa Pallotti, Sharad Singhal

In October 2016, Hewlett Packard Enterprise demonstrated the world's first Memory-Driven Computing architecture with The Machine project. This major milestone on our journey to Memory-Driven Computing is yielding technologies several orders of magnitude more powerful than any system online today. Join the people behind The Machine as we peek under the hood of the world’s first Memory-Driven Computer and learn what it takes to pull together a research and development program of this magnitude.






About the Author


Managing Editor, Hewlett Packard Labs

Rodolfo Varisco

Hello HPE fellows - with the rise of bitcoin, etheruem, litecoin, and other power-hungry blockchains, does HPE's The Machine have the ability to lower power consumption to mine these new crypto currencies?

"According to the Bitcoin Energy Consumption Index (BECI) maintained by Digiconomist, every individual Bitcoin transaction eats up 275 kWh of electricity, and the latest estimate of Bitcoin's total annual energy consumption is in the vicinity of 29.05 TWh (1)."

Very interested to hear your response to this specific use case.

Thank you!


Mr. Varisco:

Here is a response from several of our experts. 

The compute intensive Proof-of-Work consensus ensures security and scalability in public networks – requiring participant nodes to spend resources to prove that they have done the cryptographic ‘work’. The long pole of the tent when it comes to power consumption is the consensus algorithm and the silicon it runs on (CPUs, ASICs, GPUs etc.).

Memory centric architectures can certainly make the enterprise Blockchains more resilient, secure and performant. In the future, shared memory is likely to open up interesting network topologies and help answer critical problems around scale, security and energy efficiency.


But I wonder how this will be used in life science and bioinformatics. I can already see the benefits of memory-driven workload on large-scale sequencing based problems, but the crucial part in such pipelines is the asynchronous execution of modules (preprocessing, analysis and so on.). This again puts constraints on the amount of processing power available. Since in practice “The Machine” uses ARM based processors, would it be possible to concurrently run different types of modules on a single node without incurring any performance penalty - with the I/O problem already solved by memory driven architecture?

Charles Dusek

Varisco and Hopkins,

I was pleasently surprised to see Bitcoin mentioned here in the comments.  I would like to give him some context on what Gen-Z and The Machine benefits many processes like Bitcoin, but doesn't neccesarily benefit Bitcoin itself.

The Machine and the Gen-Z memory fabric enable the use of a single pool of memory by many different processors.  These processors could be the ASIC, CPU, and FPGA as mentioned before.  This sharing of memory helps more memory intensive applications and allows, for us, to place mutiple types of ASICs in the same pool.  These ASICs serve specific processes like AI, Video Transcoding and others.  While these other processes will see huge economic improvements, Bitcoin itself does not use much memory and is primarily computation bound.  Although current day Bitcoin mining would not benefit, Ethereum mining would as it is memory bandwidth bound.

And this comes with qualifications.  There may come a time when the Bitcoin block-chain is so large that memory does become a factor.  Bitcoin Cash proponents see a future when there could be 1 gb or even 1 tb blocks.  If this takes place then Gen-Z will be the go-to architecture.  Hope this helps!


Online Expert Days - 2020
Visit this forum and get the schedules for online Expert Days where you can talk to HPE product experts, R&D and support team members and get answers...
Read more
HPE at 2020 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2020
Read more
View all