Skip to content

Halloween at the lab: the measurements stay serious. Season's greetings from Alchemy Performance Lab.

MEASURE → DIAGNOSE → OPTIMIZE

Windows performance you can verify.

Remote Windows performance analysis and optimization for gaming and demanding workloads. Performance is benchmarked before and after, changes are documented, and conclusions are based only on what the measurements support.

Start with a System Check View Sample Report
  • Reversible, documented changes
  • No preset performance claims
  • You see every step of the session

WHY IT'S DIFFERENT

Measurement first, then changes.

A change is called an improvement only when repeated measurements on your own system support it. Everything else is reported as what it is.

01

Baseline first

Your system is measured before anything is changed, under repeatable conditions.

02

Controlled changes

Changes target what the baseline shows, one deliberate step at a time, never a bulk pack.

03

Re-tested the same way

The same measurements run again after the changes, under the same conditions.

04

Documented and reversible

Each relevant change is recorded, with rollback information where it can be reversed.

05

Null results count

“No measurable difference” is a valid finding and is reported as one.

06

Regressions are not hidden

If a metric got worse, the report says so.

SERVICES

Two services. One measurement method.

A System Check finds and fixes what is holding your system back. A Full Optimization tunes Windows as a whole, measured before and after. Fixed prices, paid at booking.

CREDITED TO FULL OPTIMIZATION

System Check

We measure your system, find what is holding it back and fix it. Anything that can't be done remotely comes with clear instructions.

€20

45-90 min session

Indicative; varies by system.

  • Baseline measurements under consistent conditions
  • Diagnosis across CPU, storage, system and GPU
  • Fixes for the problems found, applied remotely
  • Step-by-step instructions for anything that can't be fixed remotely
  • Baseline Report (PDF) with the findings
  • Option to re-check against a previous baseline

Book Full Optimization within 30 days and the €20 is deducted from its price.

Start with a System Check → See what's included

FULL PROTOCOL

Full Optimization

A complete, measured tune-up of Windows: we optimize the system, measure again and document every change.

€70

1-3 h session

Indicative; varies by system.

  • Measurements before and after, under the same conditions
  • Power plan and power-saving features set for performance
  • Bloatware, startup apps and unneeded background services trimmed
  • Telemetry and data-collection settings reduced
  • Clean GPU driver install and settings
  • BIOS tuning only if you want it: inside the BIOS itself, never with third-party tools
  • Verified Report (PDF) with before/after comparison and rollback list

If you see no difference, or your machine has a problem afterwards, 25% of what you paid is refunded.

Book Full Optimization → See what's included

THE VERIFIED REPORT

Not “trust me, it's faster.” A measured report.

You receive the measurements, the documented changes, and a conclusion that goes only as far as the evidence does.

Verdict per metric
Improved, regressed, within noise, inconclusive or advisory.
Before and after
Both values for every key metric, with the difference in %.
Documented changes
What was changed, why, and how to roll it back.
Measurement conditions
The measured system, the runs per side and the test conditions.
View Sample Report

Sample report with illustrative figures, not a real client's results.

Sample Verified Report page: headline verdict, methodology summary and per-metric results table Sample Verified Report page: headline verdict, methodology summary and per-metric results table

THE PROCESS

How it works

  1. 01

    Book

    Choose a service and a time, pay at booking, and answer a short questionnaire about your system and goal.

  2. 02

    Baseline

    In a live remote session that you can watch, your system is measured before anything changes.

  3. 03

    Fix and optimize

    System Check: the problems found are fixed, or you get clear instructions. Full Optimization: a complete, documented tune-up, then the same measurements again.

  4. 04

    Report

    You receive the PDF report with findings and verdicts, plus the rollback list for any changes.

WHAT GETS MEASURED

Built for games and demanding workloads

Every session combines a repeatable system benchmark suite with in-game captures when games are the workload you care about.

In-game captures

  • Average FPS
  • 1% low FPS
  • Frametime percentiles (P95 / P99)
  • Frametime spikes that averages hide

System benchmark suite

  • Storage: sequential and 4K random read, write latency
  • CPU: single- and multi-thread compute throughput
  • GPU: compute throughput and memory bandwidth
  • Timer resolution and a responsiveness indicator (advisory)

Each metric runs several times per side and is judged against its own run-to-run variation. Game results depend on the game, its settings and the scene, so they describe your workload, not a general promise. How results are judged →

WHO IT'S FOR

A good fit if

  • You get stutters or uneven frametimes despite a decent average FPS
  • Performance varies from one session to the next
  • You suspect input or system latency problems
  • You already applied tweaks and want to know whether they help
  • You want a baseline before changing anything
  • You want to know where the bottleneck actually is

WHAT WE DON'T DO

Clear limits

  • No universal tweak packs. Different hardware, software and workloads react differently.
  • No guaranteed FPS percentages. Results depend on the actual bottleneck of your system.
  • No mystery scripts. Relevant modifications are documented in the report.
  • No “more tweaks = better”. A change has to justify itself through evidence or a clear system requirement.
  • No hidden negative results. Regressions, noise and inconclusive outcomes stay in the report.

REMOTE SESSION

Your machine, your control

How the session works →

RESEARCH

Research and videos

ARTICLE

DPC latency, explained in 3 minutes

What Deferred Procedure Calls are, why they can cause audio crackle and input hitches, and what a DPC number does, and does not, tell you.

READ →

YOUTUBE

HPET Debunked - Deep Performance Analysis

WATCH →

YOUTUBE

Best Windows Power Plan for Gaming I Tested 39 for FPS & Input Lag

WATCH →

YOUTUBE

POWER TWEAKS THAT LOWER THE LATENCY | PERF STATS ANALYSIS

WATCH →
All research →

CLIENT FEEDBACK

What clients say

Excellent service. I used to spend hours tweaking Windows without always knowing if what I was changing was effective or just placebo. With Alchemy, you have a professional who has tested and benchmarked the optimizations he applies: no nonsense and no paranoid over-optimization. The system is very stable, and performance is better than ever. I finally have a usable version of Windows, and I've saved time. Alchemy is flexible, open to questions, and knows what he's talking about.
pheizy Discord
I initially came over to Alchemy thanks to his comprehensive YouTube videos that featured custom OS reviews and latency tests. […] I was first dubious because I had already had several people optimise my computer and try different operating systems, so I didn't think there was much more that could be done. However, I was fortunately mistaken, as I saw a noticeable improvement in my games following his two hours of work. This man spent two whole hours optimising my PC, when no one else took more than twenty minutes. Truly, he earned his money.
Bobby K Discord
Thanks for fixing pc latency. He fixed the stuttering in my games and the desync.
Ernis Discord
All reviews →

WHO DOES THE WORK

Vassilis

Founder · Windows Performance Analyst

Sessions are carried out by Vassilis, who runs Alchemy Performance Lab and the Alchemy Tweaks research channel on YouTube.

About the analyst →
  • Windows performance analysis
  • ETW tracing with WPR / WPA
  • Benchmark methodology
  • System optimization

FAQ

Common questions

What is the Verified Report?
It is the PDF that summarizes what was measured, what was changed and the outcome, with a verdict for each key metric. For Full Optimization it includes the before/after comparison and a rollback list. See a sample report.
Why don’t you promise preset improvement percentages?
Because every system starts from a different condition. Results depend on the hardware, Windows setup, drivers, applications, and workload. We report only what is supported by measurements from the specific system.
What access do I provide during a session?
Remote access through AnyDesk, with its unattended access enabled so the work has full access to the machine. You stay connected, can watch the entire process and can end the connection at any time. You do not need to provide passwords for personal accounts.
What if no meaningful improvement is found?
The report records exactly that. The goal is to understand the real state of the system and which changes did or did not produce a measurable result.
Can the changes be undone?
Yes. Changes made during Full Optimization are documented so they can be reversed. Where needed, additional steps or limitations are noted.
All questions →

Want to know what is actually limiting your system?

Start with a System Check or book a Full Optimization

MEASURE → DIAGNOSE → OPTIMIZE